Will New Marketing Automation Tools Let Sales Climb Back Up The Funnel?

Trends are signposts to the future: they point to where we’re headed. But the signposts are unreliable because trends are often interrupted – the classic example being the “Great Horse Manure Crisis of 1894”, when experts predicted that major cities would soon be buried beneath horse droppings. I’m beginning to suspect the much-cited trend of marketing playing a larger role deeper into the sales funnel has reached a similar peak. If the pendulum is really swinging the other way, then sales people will be taking a more active role earlier in the buying process.

This particular thought emerged quite unexpectedly during a vendor briefing yesterday. The talk had turned to the industry in general and I was ticking through my usual list of trends – external data, predictive modeling, sales enablement, advanced attribution, adtech and martech integration, local/partner marketing systems, all-in-one systems, content creation support, and of course customer data platforms. But I also had in mind another system I had just seen, MDCDOT, which gives marketing automation functions to sales people. For some reason, I suddenly saw a connection of this to external data: if vendors like NetProspex and InsideView could provide sales people with prospects and vendors like MDC provide tools to nurture those prospects, then sales people really don’t need marketing to do either of those things. Sales could then push marketing back to its traditional narrow role of generating promotional materials and doing research. That's pendulum swinging with a vengeance.

Once an idea like that pops up, supporting arguments quickly fall into place.  Integration of advertising with marketing technology potentially gives sales people another route for generating their own prospects. Advanced data enhancement and lead scoring make it easier for sales people to automate lead nurture processes without becoming marketing automation experts. All-in-one systems and customer data platforms both unify marketing and CRM technologies, making it easier to shift boundaries between marketing and sales responsibilities.  Those shifts could be permanent or vary dynamically based on fluctuations in needs, resources, and individual interests. Sales enablement tools like Velocify and Clari help salespeople pick the right treatment for each prospect, which can include sending them to automated campaigns.

In short, the conditions may be ripe for a counter-revolution: a sort of Terminator 2-style conversion where some machines defend the humans instead of trying to replace them. It even dawns on me that having marketers restrict their focus to content creation has some advantages, given how much more content is needed these days.

Let me be clear: this is a shiny new idea which could quickly lose its luster. It only applies in B2B and considered-purchase B2C relationships where actual salespeople are involved in the buying process. And it may be that most salespeople are happy to let marketers handle the lead generation and nurturing, which they never enjoyed in the first place. Or perhaps corporate management will decide it’s more effective and safer for marketing to handle those tasks, regardless of what sales people want.

On the other hand, sales departments are usually much larger, more powerful, and better funded than marketing departments, so it could be in the marketing automation vendors’ interest to serve sales people directly. This leads to the long-expected assimilation of marketing automation into CRM, a trend that has never quite happened but might finally take place.

Or maybe the future is really about machines selling to other machines, and the whole distinction between sales and marketing will no longer matter. Only time will tell. In the meantime, bear in mind that you never know whether a signpost is correct until you’ve passed it.*
__________________________________________________________________________________________
*I don’t know what that means, either, but it sounds pretty deep, don’t you think? I can definitely see the motivational poster.

Ensighten Transforms Web Tags into Rich Customer Data

Looking back once more at last month’s post on the future of marketing data, you may recall that I briefly mentioned the intriguing rise of Web tag management systems as platforms to integrate customer data. Tealium highlighted the topic on Tuesday with $30.7 million in new funding, bringing the total to $77.9 million to support its stated mission of “capturing and serving up real-time, 360-degree customer insight to all of their [customers’] enterprise applications.” Sounds like a Customer Data Platform to me.

But I'll get back to Tealium some other time.  Today I'm writing about Ensighten, another major player in the tag management space, and one that is also pursuing a much grander vision: to be an “open marketing platform” that lets clients build a “custom marketing cloud”.

If origin stories count, and I think they do, then it’s worth noting that Ensighten’s started in 2009 as a digital analytics agency. Frustrated by the limits that traditional tagging methods placed on clients’ ability to move quickly, the firm developed an engine to generate custom sets of tags for each Web site visitor. This differs from the more common approach of serving one container that holds all possible tags.  (The problem both solutions address is that many Web pages contain a large number of tags from different systems; tag managers consolidate those tags so load more efficiently and are easier to control.) Ensighten’s method meant that only the tags that would actually fire for a particular visitor were loaded, while a container serves all tags and lets the Web page figure out which to fire. Ensighten studies have found it reduces page load time by 40% and tag deployment time by 75%.

Whatever the relative merits of Ensighten’s approach to tag management, the company’s focus today is on the data itself.  In particular, Ensighten wants to ensure data from the tags is available for the client’s own use. This regains control that is lost when partners place their own tags on company Web pages and don’t necessarily share the resulting information.

As I’m sure you see coming, Ensighten’s next step is to assemble the tag data into detailed individual profiles for personalization, testing, attribution, and other purposes. Ensighten added these capabilites a year ago, just after taking a $40 million investment (bringing its own total to $55.5 million if you're keeping score). These features go beyond Web page tags to capture data from mobile apps, from ad pixels served on external Web sites, and from other systems via batch data imports. They also include processes to merge profiles using customer IDs, cross-reference tables for IDs from different systems, probabilistic matching of shared data elements, correlated behaviors, and links between individuals and devices.  This kind of identity association is arguably the hardest part of building an effective customer database and many systems don't do it.

Ensighten can deliver rule-driven personalized messages along with the custom page tags. It also provides services including Web site performance monitoring and privacy management and will apply machine learning to personalization and recommendations later this year.  But such applications are peripheral to the company’s long-term strategy of letting external systems use the data for their own purposes.  This is what puts the  “open” in “open marketing platform” and the “custom” in “custom marketing cloud”.

This approach means the company’s real competitors are other marketing suites and customer data management systems. Since last year’s funding, Ensighten has supported its strategy with acquisitions of tag management competitor TagMan in March 2014 and cloud analytics and predictive marketing vendor Anametrix in October 2014.  The TagMan deal put Ensighten's foot inside many new corporate doors, while Anametrix provided advanced technology for customer database management and analytics.

So far, it all seems to be working. Ensighten reported 150% revenue growth last year and says it now has “a few hundred” global clients, mostly very large enterprises. Pricing is based on volume, which could be Web traffic or audience size depending on the situation. The company doesn't release details but this is enteprise software: you can safely assume it’s not cheap.

I’ll end where I started: what Ensighten (and Tealium and others) are doing is remarkably clever: take a mundane service (tag management) that places them in a strategic position (touching all digital interactions) and use it to build a strategic service. Their starting point gives them a shorter path to building a company’s primary customer database than applications like lead scoring or customer success management, which must create new data flows before they can build the database they need to support their service.  The tag management vendors' position also lets them give equal access to all applications without worrying about competition with their own application services. And, from a purely practical stand point, it also gives them an initial relationship with many potential clients. This advantage alone is probably why Ensighten has so many more clients than the “pure play” CDP vendors Aginity or NGData I wrote about earlier this week.

Of course, the tag managers' success is not guaranteed. Building an integrated, persistent customer database is quite different from managing Web tags.  But it's always a good thing when marketers have more options.  Companies looking for help in building their core customer database should definitely take a look in the tag managers’ direction.

NGData Gives Enterprise Marketers a Customer Data Platform of Their Own

If you read my recent post on Customer Data Platforms Revisited very, very closely, you might have noticed it listed a category of data vendors who “store unified profiles and expose to other systems”, which is pretty much the core definition of a Customer Data Platform. You would also have noticed that category had only two members, Aginity and NGData. I reviewed Agnity back in November 2013 and when I spoke with them more recently, found they were still doing pretty much the same thing and growing nicely. But, until today, I’ve never discussed NGData.

I wish I could come up with a really cool reason for that, but it’s only because NGData just recently came to my attention. The company itself has been serving clients in late 2012. As my classification suggests, they are in the business of assembling client data from multiple systems, using it to build detailed customer profiles, and making the results available to execution systems like campaign management, call centers, Web sites, and mobile apps.

The technology involved is Hadoop as the primary data store and HBase to expose the profiles to external systems. Data is unified mostly with customer IDs supplied by the source systems, although NGData can also build cross reference tables to associate related IDs and do some probabilistic matching to link related devices. The profiles include both raw data and calculated metrics such as trends, exceptions, signals, affinities, predictions of fraud risk, churn likelihood, and next product to buy.  The predictions can be based on conventional predictive methods like regression or on integrated machine learning. Often the company is able to use regression models that the client has built already for other data sources. The base version of the company’s system, called Lily Enterprise, has about 700 such metrics, and clients can add their own. (There’s also an open source version of Lily, which has only the data management components of Lily Enterprise.)  External systems access profiles via SQL queries against HBase, API calls, or file transfers.

This may all sound pretty straightforward and should be familiar to readers of this blog. But out there is the real world, most companies are still struggling with the challenges of assembling customer data and making it accessible.  Those firms will find this is pretty novel stuff. NGData stresses its ability to easily add new data sources and to retain huge amounts of detail, something it inherits from Hadoop. Complex metrics like exceptions and trends are also relatively unusual.  They make it much easier for execution systems to act intelligently, since the hard work of surfacing opportunities and selecting responses is largely done in advance.  Predefining those metrics as part of the base system speeds initial deployment.   One proof point: NGData says clients can typically start running programs based on its system in four to five months, and sometimes as quickly as three months. Conventional data warehouse projects usually take more than one year and many are never delivered.*

NGData stresses that marketers get better results when they can look at all the data associated with each individual, rather than treating people as members of broad segments. Again, no reader of this blog will disagree.  But I did still like their example of using behaviors to identify people who are likely to call customer support with a particular problem, and then preempting those calls with personalized messages about how to solve the problem. Everybody wins: the customer appreciates the proactive service, and company fields fewer phone calls. The underlying point, which NGData stresses often, is that they're going beyond insight to producing an actionable result.

NGData also highlights the value of responding to customer behaviors in real time or near real time.  It can do this because it immediately updates its metrics in Hbase whenever it receives new information. Still is more old news, but conventional solutions often update their scores and recommendations nightly or even less often. To be fair, NGData can only update in real time if the source systems are providing immediate updates – which often isn't the case.  The vendor does have its own Web tags to capture real time information about Web visits, which lets it use in-session behavior to drive product and page recommendations.

The one thing you may not expect about Lily is that it’s old-style on-premise software, not software-as-a-service. That’s mostly because current clients are big companies in financial services, telecommunications, and media, industries that have been reluctant to let precious company data outside their walls (although plenty of hackers still manage to get it, he added snarkily). It’s also not clear how much SaaS would benefit NGData, since each client’s data store would remain separate in any case and the software needs to be installed on relatively few desktops. For what it's worth, Aginity is also on-premise and also serves primarily enterprise clients. 

I suspect it's no coincidence that the two "pure" Customer Data Platform vendors both focus on enterprises.  Smaller firms need to justify a CDP by combining it with a specific application, but big companies can afford a separate CDP project that they'll later tie to separate execution systems.  Perhaps this will change as CDPs get cheaper, the model is better understood, and integration with execution systems becomes easier.

NGData has about 25 current paying clients in Europe and the U.S. Pricing is based on volume of data and/or number of customers. Total cost for software and services starts around $150,000 to $200,000 and can go much higher. As I said, this is enterprise software.

________________________________________________________________________________
*This is one of those things that “everybody knows” but few cab document. Here’s one paper that says data warehouses “usually” take 12 to 36 months, cost $1 to $1.5 million, and have a 70% failure rate. It’s not clear where the author got his data but it all sounds about right.

VEST Report: Latest Trends in Marketing Automation, and Where's My Hoverboard?

I just finished the latest release of the B2B Marketing Automation Vendor Selection Tool, a.k.a. VEST Report. The new version includes a big technical change: instead of the interactive Flash document that was very cool but people often had trouble running, it’s now a combination of PDF for the core document and Excel spreadsheet for the detailed vendor scores. That’s a technical step backwards but will actually make it easier for buyers to access the detailed vendor information, and in particular to screen for vendors with particular capabilities. Less is more, I suppose. The good news is that this format lets me expand beyond 25 vendors, which was the maximum the old system allowed before running out of memory.

Of course, none of this is your concern, Dear Reader. What’s you'll find more interesting is that the VEST provides an opportunity to see new patterns emerging in the industry. Usually I do this by taking a close look at which features have become more common since the last report. But this time there were a few more obvious changes that stood out. Here’s what struck me.

- more micro-business vendors. All six of the vendors new to this report sell primarily to small businesses, and most are “all-in-one” systems that combine marketing automation with integrated CRM. They join another six vendors from previous editions who also serve this market. I'm also aware of several other vendors, not yet in the VEST, who also compete for this business. Many of these firms are new while others have been around for a few years but just hit my radar. What this says to me is that the all-in-one segment is more crowded and more mature than it has seemed. Of course, there’s still a huge opportunity – hundreds of thousands if not millions of potential clients have yet to buy their first system. But anyone planning to enter this business had better realize they will be fighting for new customers.*

- agency relationships. It seems that just about every vendor in the VEST now touts special features to support marketing agencies that resell the system to their clients or operate the system on the clients’ behalf. This isn’t exactly new but what once seemed like a niche strategy now looks more like a standard approach. It’s always been obvious that agencies were a sensible channel for marketing automation vendors to pursue, but I’m beginning to wonder whether agencies might turn out to be the primary channel for such systems, excepting only direct sales to large enterprises. If this happens, the reason will be that agencies provide the missing skills that have prevented so many companies from taking full advantage of marketing automation systems by themselves. Vendors have been knocking themselves out for the past five years trying to educate marketers to run their systems.  Perhaps having agencies run them is the real solution instead.

- social data. Maybe my biggest surprise was finding that many if not most vendors have now added features to automatically look up new contacts in social networks and add that data to their marketing automation or CRM profile. This seemed like magic three years ago when I first saw John Ferrara's Nimble do it; but now it’s commonplace. In fact, any vendor that hasn’t developed their own technology can just integrate FullContact to do it for them. So the competitive advantage is now precisely zero. (Okay, not zero: some companies surely do it better than others.  But that’s a much weaker selling point than being one of the few firms to do it at all.)

- ad tech integration. This one isn’t so common yet, although Oracle Eloqua, Marketo, HubSpot and some others have announced some ad retargeting partnerships. Google Adwords integration and advertising through Facebook, Twitter, and LinkedIn audiences are more widely available but I don’t include them here. But despite the slow growth, there’s no question that serious integration between Web display ads and marketing automation programs will become much more widely available. What I won't do is predict how quickly that will happen. But I’ll certainly add it to the list of VEST questions so I can track it more closely in the future.

- dogs that didn’t bark. That’s a Sherlock Holmes reference, not an insult to technologies that haven’t been as widely adopted as the industry seemed to expect. Okay, maybe it’s a bit of both. In any event, I didn’t commute to work today on my hoverboard, and you probably didn’t sit down to do advanced mobile marketing, predictive modeling or revenue analytics in your marketing automation system. Those three – mobile, predictive, and revenue analytics – are all technologies that should take off, but so far are not deeply integrated with most marketing automation platforms. Maybe mobile has become so ubiquitous that I don’t even notice it, but, so near as I can tell, few vendors have done more than make it easier to create emails and Web pages that look good on mobile devices. Surely mobile can do more than that. Predictive analytics are growing quickly but so far are still done by specialized vendors rather than built into the marketing automation platform. (Yes, there are a few exceptions like the machine learning features of dbSignals and RedPoint. But they’re exceptions.) Revenue analytics is only discussed by a couple of companies; although important, it doesn’t seem to have captured the industry’s imagination. I haven’t given up hope for any of these, but no longer expect them to quickly become part of the mainstream.

So those are my impressions while the VEST updates are fresh in mind. The report is well worth buying if you want do to your own industry analysis, or (its primary purpose) are searching for a new system. As I say, the new format does make finding vendors with specific features much easier. You can find more information or place an order at www.raabguide.com/vest. Let me know what you think.


__________________________________________________________________________________
*In fact, the micro-business segment is even more complicated than I’ve suggested. The real competitors are companies like ConstantContact who are providing a broad range of services, such as local advertising, that extend well beyond marketing automation and CRM. There are also many vendors with specialized services for vertical markets such as plumbers and funeral homes. Come to think of it, I recently noticed that one of my many plumbers (don’t ask) uses a system developed by a funeral home website firm. If there’s a logical connection between those businesses, I don’t want to know about it.

New Marketing Automation Options for Small Business in the VEST Report

I’m revving up for the next edition of our B2B Marketing Automation Vendor Selection Tool (VEST) report, which will include six first-time entries. I’ve already written about two of those, Inbox25 and AutopilotHQ (formerly Bislr). Here are thumbnails of the others.

GreenRope workflow

GreenRope is all-in-one software sold primarily to very small businesses such as lawyers, real estate agents, consultant, coaches and membership organizations. That puts it firmly in Infusionsoft territory, and perhaps even towards the lower end of that. The system has an impressively broad scope, adding full Web site creation to the usual all-in-one mix of email, lead scoring, landing pages, and CRM. The features within these functions are unusually sophisticated for a micro-business system: email includes dynamic content and a/b tests; Web pages also support a/b testing; Web forms allow progressive profiling; email and Web responses can automatically trigger a follow-up action.  CRM includes opportunity tracking, unlimited user defined fields, and automatic search of Facebook and LinkedIn when new contacts are added; algorithms can automatically estimate the right number of points for different events to build a predictive lead score.  The media library supports images, files, articles, and videos (through Vimeo integration); the calendar provides full event management; surveys can change questions based on previous answers and include automated follow-up actions.


The system also extends beyond sales and marketing functions to include customer forums, Wikis, support tickets, project schedules with tasks assigned to individuals, and coupons. Workflows can manage both marketing campaigns and internal projects. Additional functions are provided through integration with other systems, including Olark for chat, Twilio for voice and text messages, VoiceBase for transcription, Magento for ecommerce, Quickbooks for accounting, and Microsoft Outlook for email.

In other words, although GreenRope describes itself as “CRM and marketing automation,” it actually extends beyond those functions to manage activities throughout the business. This is very desirable for small organizations that want to automate their operations while running as few systems as possible.

GreenRope is also small-business-friendly, starting at $149 per month for up to 1,000 contacts and costing $199 per month for 5,000 contacts.  All plans include unlimited users and unlimited emails, which isn’t always the case with small business systems.

While GreenRope is new to the VEST report, the company itself was founded in 2008.  It currently has about 4,500 end users at a somewhat smaller number of companies.

Hatchbuck workflow

Hatchbuck is another all-in-one product for very small businesses. It has taken the approach of providing only core features and making them as easy to use as possible. This scope covers email, Web forms, multi-step campaign flows, and CRM. The system integrates via Zapier with ecommerce products. It recently added lead scoring and the ability to look up individuals on social networks. The company serves a mix of clients, with the largest segments including technology and manufacturing companies, travel, and professional services. Most clients have fewer than ten employees.

As the company’s strategy suggests, Hatchbuck provides basic capabilities for its core features but skips the more advanced options. It creates email templates with personalization and embedded links but no dynamic content. CRM captures activity history, tasks, deals, purchases and events but doesn’t integrate with a phone dialer. Forms can be associated with actions but there is no specialized survey builder. You get the idea. Instead of adding more features, Hatchbuck’s developers rigorously benchmark the number of clicks it takes to perform system functions in Hatchbuck and competitive products and track how customers use each feature to identify problems. The company also provides extensive training and support materials, including a required three hour Quickstart package to help new clients use the system effectively.

Hatchbuck was founded in 2011 and launched its product in 2013. It now has about 700 customers with over 2,000 end-users. Pricing starts at $99 per month for one user and 2,500 contacts and reaches $199 per month for three users and 10,000 contacts. All plans include unlimited emails. The Quickstart package costs $199 but the fee is waived for clients who sign a six month contract. The company also has special features for marketing agencies who use the system for their clients. Such agencies account for about one quarter of the Hatchbuck business.

Lead Liaison content creation

Lead Liaison calls itself “revenue generation software” to indicate that it provides more than a standard B2B marketing automation product. Additional features include lead distribution and buying signal alerts, but don’t extend to full CRM or the other operational functions. With a $500 per month starting price, it is targeted at small to mid-size businesses but not at the most tiny. Pricing is based on the number of contacts in the database, with unlimited users, emails, and page views.  The company doesn’t publicly state how many contacts that $500 gets you.

The system offers advanced versions of the usual marketing automation functions: email, landing pages, Web forms and surveys, lead scoring, multi-step nurture flows, media hosting, and CRM integration with Salesforce.com, Microsoft Dynamics CRM, and Sugar CRM. It also goes beyond these in several directions, including:

- company-level Web visitor identification based on IP address, which can be tied to Data.com or LinkedIn to pull back the names of individual contacts at the identified companies (although these are not necessarily the actual visitors).

- matching of contacts against social networks to add their social identifiers to the Lead Liaison record

- phone dialer with scripts, call notes, and a payment widget

- option to send emails from LeadLiaison’s own servers or through third party services including Mandrill, SendGrid, and SMTP Inc.

- social media posting to Facebook, LinkedIn, and Twitter, including an option to store posts in a queue that will release them on a regular schedule

- a nifty Web page scanner that can copy an existing Web page or form from any source into a version that the marketing automation user can edit by, say, inserting a Lead Liasison form or link

- agency-friendly features including single log-in to multiple accounts.

Perhaps the most interesting feature of LeadLiaison is a content creation wizard that connects to a network of prequalified writers for blog posts, white papers, press releases, newsletters, Web pages, social media posts, and other materials. This is directly integrated with the system: users fill out a form specifying their requirements, which Lead Liaison submits to the network.  Once a writer (whose identity is hidden from the user) accepts the project, the system tracks the material through production states and eventually loads it into the Lead Liaison asset library.  Assets are automatically coded so users can track consumption.  The system can also limit distribution based on date range, number of downloads, and whether visitors are asked or required to provide an email address to receive it. Pricing is modest: a blog post costs $50 with five day turnaround. Although the writers are anonymous, LeadLiaison plans to let users favor authors of specific pieces for future assignments.

LeadLiasison was launched in 2013. It has under 200 clients and serves a mix of B2B and B2C marketers.

dbSignals workflow

dbSignals is brand new: the system was formally launched just last week. (Full disclosure: I’ve consulted for them.) The system straddles B2B and B2C marketing automation, using a flexible data structure typical of B2C products but also providing Salesforce.com integration, the B2B hallmark. It also includes its own lightweight CRM.

The marketing automation functions themselves are quite sophisticated: dynamic content, multi-step branching campaign flows, multivariate testing, fine-grained user rights management, option to use internal or external email services, and integration with external HTML templates. Supported channels include email, SMS, direct mail, surveys, landing pages, and social media. There are also options to support marketing agency users, including an ability to rebrand the system with the agency or client’s own identity. 

And, yes, the system also can look up the social profiles of individual contacts and add them to its database.  That feature has quickly become a new standard.

But what really distinguishes dbSignals are two features beyond the normal scope of marketing automation. The first is prospect data: the company has negotiated deals to let its clients access detailed files with 235 million consumer names and 60 million B2B names. These are selectable within the normal system interface, along with whatever names a client loads on its own.

The second feature is machine learning.  This is initially being deployed to identify the most responsive list segments within the prospect data. The process is wholly automated: the only choice users make is whether to turn it on.  Once they do, the system analyzes the client's customer list or past campaigns, builds a predictive model, runs test campaigns to validate and refine the model, and then runs a roll-out campaign once the model is stable. Models are further adjusted after later campaigns. dbSignals will soon add other uses for machine learning including churn prediction, lifetime value prediction, and attribution of the incremental impact of marketing programs.

Prospect data and machine learning are closely integrated.  Indeed, one of the reasons machine learning can be so fully automated is that the system can rely on the prospect data elements to be available -- including up to 2,000 variables on a consumer profile. Beyond that, dbSignals uses the machine learning results to “reserve” the best prospect names for each client in advance of campaign selection.  This is needed because dbSignals limits the number of promotions sent to any name within a specified time period.

Both the prospect data and machine learning are in turn made possible by dbSignals' underlying technology, which uses the Cassandra data store instead of a standard relational database.  Few marketers will care, but, trust me, it really matters for speed, scale, and flexibility.

dbSignals also offers an unusual pricing model, basing charges on the number of users and/or message volume rather than database size. This makes it easier for clients to take full use of the prospect data. Fees start as low as $500 per month.

The initial version of dbSignals was introduced in 2014. The company currently has about two dozen clients including a mix of B2B and B2C organizations.

Customer Data Platforms Revisited: The Future of Marketing Data

It’s nearly two years since I introduced the concept of a Customer Data Platform, defined as a marketer-controlled system that builds a multi-source customer database and exposes it to external execution systems.  You may recall that I listed several sets of products as CDPs: B2B predictive lead scoring and customer success management; campaign management with an integrated customer database; and data management platforms to support online advertising. Systems were included only if their data (or derived data such as model scores) was available to other systems for campaigns and messaging.

All those categories have done well since my original posts on the topic. Established vendors have grown quickly and attracted funding; new vendors have joined the mix, also often with substantial funding. So I suppose I could pat myself on the back for spotting an important trend and let it go at that.

But things aren’t quite so simple. A look at the entire CDP ecosystem uncovers important patterns that are hidden when you look at individual vendors or vendor categories. Here's a summary of what I've seen.

Customer Management Functions

CDPs exist because marketers need to coordinate customer (and prospect) interactions across channels. That coordination involves three basic tasks: gathering and unifying customer data from all sources; using that data to select the best treatment for each interaction; and delivering those treatments through the appropriate channel systems. Each of those three tasks has several subtasks. These layers are illustrated by the following diagram, which includes a unified data layer – the classic CDP.

Vendor Categories

So far so good, but it’s really just theory. Things get interesting when you look for specific systems that perform the subtasks. It turns out that there are several categories of specialist systems within each subtask, each doing similar or complementary things in slightly different ways. Connecting the logical flow to actual systems is important because looking at real products tells you what the market is saying: that is, what buyers are willing to pay for and where change is concentrated.

The following table shows what I found when I did this analysis. The list of vendors in each section isn't necessarily comprehensive, especially in crowded segments like B2B marketing automation. I should also stress that I’ve only included Decision-layer vendors who also build their own database. This makes them potential CDPs and means they have many Data-layer functions. In a sublimely liberating act of inconsistency, I have NOT limited the Delivery layer to vendors who build their own database. In fact, most do not.

Investment

The right-most column on the previous table shows the level and types of investment being made in each vendor class. I haven’t collected precise details but the general patterns are pretty strong. The major observation is that current investment is heavily concentrated on the Decision layer, with interest in predictive modeling and message selection (which could also be labeled as personalization). There’s some investment on the Data layer in data gathering vendors, especially along the lines of acquisitions by big companies (Oracle/Datalogix, D&B/NetProspex, etc.). This is a general sign of maturity. Similarly, most recent investment on the Delivery layer has been acquisitions (IBM/Silverpop, Oracle/Responsys, Teradata/Appoxee, etc.), which is a sharp contrast from the heavy venture capital funding a couple of years back. Again, this shows the relative maturity of the space.

(Caveats: although it doesn’t show up in this analysis, I do still see some interesting investment in marketing automation niches such as app marketing, distributed marketing, and agency systems. I’m also increasingly intrigued at the “tag management” vendors on the Data layer (Tealium, Signal, Ensighten, etc.), which are reinventing themselves as data integration hubs. I didn’t see that one coming.)

Implications

It’s tempting to interpret these results are showing that data assembly is a solved problem, allowing marketers to invest Decision systems on the next layer down. But any marketer can tell you, and every survey I’ve seen confirms, that most companies are nowhere near having fully integrated their customer data.

What I think is really going on is that people are investing in Decision systems that build their own multi-source databases, providing both Data and Decision functions in one package. Remember that my original CDP categories included B2B predictive vendors and campaign management vendors who did exactly that. So it seems the proper way to look at things is more along the lines of the following diagram, which shows there are several different ways to solve the customer data integration challenge: you can buy a stand-alone CDP that has only data-level functions; buy a Decision system that also builds an integrated database; or buy a Delivery system that does data, decisions, and execution. As the diagram indicates, most of the Decision vendors do incorporate the CDP functions, while only a few of the Delivery vendors do.

The diagram labels the Data + Decision combination as a “Marketing Platform”.  I think this is reasonably consistent with how most people use the term, since the key feature of a “platform” is its ability to integrate with external systems for delivery and other purposes. I’ve labeled the Data + Decision + Delivery combination as an “Integrated Suite” and used question marks to show that not all suites provide a complete Data solution. This is because many suites aren’t very good at bringing in external data or letting external systems access the data they’ve assembled.

As I noted in the previous section, most of the industry funding and excitement is centered on the Decision layer, which is where the Marketing Platforms live. The practical advantage of those systems over Data-only solutions is obvious: Decision systems deliver a revenue generating application while Data-only systems do not.

But think about that for a moment.  Each Decision system builds its own multi-source database and each integrates separately with the Delivery systems.  Having multiple Decision systems is a nightmare of redundancy:

It seems pretty clear that the better solution is to have a single Decision system controlling everything, which is arguably what most people (and vendors) have in mind when they describe a Marketing Platform. Indeed, this is exactly the direction that most Decision-layer CDPs are headed, by expanding the scope of their products from an initial point solution, such as B2B lead scoring, to encompass other applications. It’s safe to say that the people who built these systems always planned, or at least hoped, to grow in this direction.

Does the growth of Decision-layer CDPs mean that Data-only CDPs will fail? I’ll admit that only a few such systems have appeared in the past two years. But I’m not quite ready to give up on the concept.

Why?  Well, as Tolstoy never said, all good customer databases look alike, but every decision system is different. This means it’s hard to support all types of decisions within a single product. So it does seem that multiple decision systems will appeal to marketers who have the skills to use them and the scale to justify the added expense. Those marketers would benefit from a Data-layer CDP, which would make it easier to deploy best-of-breed decision tools even when those tools lack data unification functions.

The stumbling block for this approach is still the cost of integrating multiple systems: as the diagram shows, there are still plenty of connections in this model. But there’s at least some hope (although I remain skeptical) that newer technologies will make the integration easier. The other bright spot for the Data-only CDPs is that they should be attractive as partners or acquisitions for Decision and Delivery systems that haven’t built their own CDP functions.


And what about the suites? I’ve said for years that the first law of software market development is “suites win”, precisely because most companies will sacrifice best of breed functionality to avoid the costs of integration. Indeed, the big marketing clouds from Oracle, Salesforce.com, IBM, Adobe, and others all include extensive Delivery layer functions. I think it’s fair to say that while their commitment to being “open platforms” is genuine, they see that as a way of letting clients supplement the core functions the suites provide internally.  This is quite different from the idea of a shared Data and Decision platform that specifically avoids offering Delivery services. Still, there’s a  very good chance that a suite which can easily integrate supplementary functions will give marketers enough freedom to overcome the problems of lock-in, while still delivering the convenience of pre-integrated core functions. So I’m not quite ready to abandon “suites win” as a rule, although I’m a bit less certain than previously.

Looking Ahead

It’s fun to handicap the horse race among vendors and categories, but what really matters is the contest itself.   All these smart people and money are finally giving marketers the unified customer databases they so desperately need.  This removes a fundamental obstacle to the cross-channel integrated marketing that everyone recognizes is increasingly important. So let’s look at the view once we've climbed that mountain.

I’d like to tell you I see a new and perfect world, but what's actually there is more mountains.  Once unified databases become available, marketers will face a new set of challenges including:

- more need for predictive models and external data. I only lump those together because they’re already getting a lot of attention. Having a powerful database just makes them even more important.

- new focus on automated content creation and campaign design. Lack of skilled users and adequate content are already huge barriers to effective multi-channel marketing.  Removing the database barrier will only make them stand out even more. So we can expect smart people to address them through technology. Indeed, there is already plenty of activity in these areas but I think it’s fair to say that so far none of vendors have had a major impact. This is arguably the next exciting frontier for marketing technology.

- more developments in cross-channel customer tracking. Again, the need for this has been obvious and some major investments have already been made. Cookies are becoming increasingly inadequate as cookie-hostile channels like mobile become more important. Marketers will soon reach a tipping point (or maybe they already have) where they realize they must abandon cookies and move on to other approaches such as device identification or external identity databases. A new standard will eventually emerge, although I can’t even guess what it might be.

- tighter integration between advertising and marketing technology. These two realms are now largely separate with a few exceptions such as retargeting. But as personalized ad messages become increasingly possible, marketers will have ever-greater incentive to target and, ultimately, coordinate messages across channels using shared data. This is highly dependent on the improved customer tracking, so it might have to wait a bit.

- better marketing attribution. If there’s a last stop on the road to marketing Nirvana, attribution might be it. Once marketers have assembled all that data and associated everything with the right customer, they’ll finally be able to deploy advanced analytical methods to really understand the long- and short-term incremental impact of their marketing efforts. Then, and this itself would be heavenly, we’ll never again hear anyone quote John Wanamaker about not knowing which half of his advertising is wasted.

Recommendations for Marketers

Nirvana is still far distant.  Marketers face immediate choices in how to spend their time and budgets. The trends I’ve just described do have some immediate practical implications. Here are my suggestions:

- Experiment like crazy. The various Decision-layer vendors currently offer different specialties, such as lead scoring vs. product recommendations vs. churn predictions. Vendors in each area are expanding their scope so there’s a good chance you’ll eventually pick one to do almost everything. To have the best odds of making a good selection, you’ll want to learn about as many vendors as possible in advance. So run tests to build an understanding of the applications, technologies, and corporate culture. The good news is that each approach can probably pay for itself in improved performance, so these tests should be more or less self-financing.

- Keep an eye out for new data. Many of the Decision-layer vendors bring their own data to the party, and evaluating that data is one part of understanding what they offer. But there are also other data sources that are not tied to a Decision system. You’ll want to explore these to understand what value they provide value and whether to make them part of your long-term data foundation.

- Plan for integration. You may not have shared customer data or decisions today, but it’s increasingly likely they’re in your future. So every new marketing system should be evaluated in part on its ability to integrate with other systems. This involves sending data to the central database and reading data from it, as well as integrating with Decision-layer systems for predictive models, rules-based selections, optimization, recommendations, personalization, and more. Even if you’re going to use an integrated suite, you’ll want to assess how easily you can supplement its functions by tying into external products, and what kinds of products are already available for integration.

Summary

The stand-alone Customer Data Platform is one solution to the challenge of providing a multi-source, shared marketing database, but it isn't the only option.  Whichever solution marketers ultimately find most appealing, they will benefit from gaining control of their data and moving on to new opportunities that database makes possible..

More on Marketing to Things

I’ve been working on that screenplay about marketing to things (see Do Self-Driving Cars Pick Their Own Gas Station?). It’s not going well – all the scenarios lead to self-aware computers taking over the world, which is both depressing and unoriginal. But I did come up with some interesting thoughts to consider while you’re waiting for that (computer-controlled) ball to drop at midnight. In no particular order:

- message overload is a fundamental problem for marketers: people get so many messages that it’s increasingly difficult to break through the clutter. Marketing directly to machines offers a way avoid the overload, especially as machines take over more of our day-to-day decision making.

- skeptics might think that machines will just buy the things people tell them, so they make no choices and therefore there’s no way to market to them. But most agree that they’ll want machines to have some discretion such as how far to travel for the lowest price gasoline or how to balance cost vs. quality when selecting hotel rooms. Once a machine starts balancing different factors, marketing opportunities arise: for example, a car that's told to minimize total operating costs might choose more expensive gasoline if that’s combined with a discount on an oil change it will need the near future. The trick will be to understand the algorithms that machines use to calculate value and to offer the most attractive bundles taking into account value, price, and actual cost. If that’s not marketing, what is?

- “Know your machine” may become even more important than “know your customer”. Sticking with cars, they could easily expose their maintenance and performance records to nearby merchants via short range wireless.  Merchants could then compete to offer the most compelling package of goods and services based on each vehicle’s current condition. Imagine your car cruising down the street being solicited by every gas stations it passes . It gives a whole new meaning to the term, “red light district”.

- machines may interfere with other machines along the path to an actual purchase. Imagine that your exercise wristband recommends healthy recipes and asks your grocery program to buy the ingredients.  But the grocery program makes adjustments based on in-store specials, and your refrigerator then vetoes certain items because it knows you already have them (or, cleverer still, because it knows you always throw them out unused). Once this sort of thing starts to happen, there’s an opportunity for the different machines to negotiate with each other or for suppliers to incent the different machines to favor their products. And there’s really nothing to require that those incentives all be paid to the consumer.  Retailer rebates are part of the business world today and this would logically be the same thing.

- there may be certain arenas that are the site of intense competition, in the way that bankers compete for “share of wallet” and restaurant chains compete for “share of stomach”. I’m thinking food choices may be one option – think “share of refrigerator” or “share of cupboard”. Purchases for you automobile are another focus; so are travel choices. In each case, expect competition to be the gatekeeper: a master coordinator that makes the final choices and, as noted above, might be able to charge access fees to everyone else.

- devices might try to modify your behavior, presumably for your own benefit. Imagine that the black box in your car (which is already there today) notices you’re a safe driver and that your auto insurance company offers a discount for such behavior if you agree to be monitored. Wouldn’t you want the black box to tell you about the opportunity since you already qualify? Or maybe that fitness program is watching your calorie intake as part of a health insurance incentive plan: when it notices you’re about to go over the daily limit, it suggests a gym session tomorrow morning to bring things back in line and automatically adds it to your calendar. Or, more elaborately, the fitness program sees from your calendar that you’re about to make a reservation at a fattening restaurant: it warns you and suggests an alternative. If you really want to get scared, think of the fitness program as automatically tracking your calorie intake by tracking what you take from the refrigerator via RFID, what you buy at vending machines via charge records, where you’re eating via GPS, and what you order via voice recognition. I guess it would be easy to cheat and not all that accurate, but how accurate are the current self-reported data that most people use to track their food intake?

- calendar programs could become especially important. It’s easy to imagine telling your calendar to plan a dinner date tomorrow with someone, and then having that calendar talk to her calendar and find a good spot based on mutually convenient locations, where you’ve both eaten recently, food preferences, personal ratings for known restaurants, published ratings for new ones, and maybe what kind of mood it predicts you’ll be in given the events on your calendar that afternoon. If the schedule looks really ugly, it might just switch the location to a bar and pre-order your martinis.  If multiple people are involved, their calendars might even negotiate the date itself.

- merchant ratings will be increasingly important currency as machines make more decisions and the decisions are driven by ratings. Today, many companies ask every customer to submit a rating. But a clever program could deduce satisfaction from things like size of the tip, on-time arrival, problem resolution time, below-average repair costs, and other variables that depend on the product. It would then ask only the best-served customers to make ratings. A really smart system would offer dissatisfied customers an apology or incentive to come back. The system could even flag those customers for special attention on their return visit. (Ok, this isn’t directly “marketing to things”, but the data is coming from “things” that have been instrumented.  Plus, if a system like this doesn't already exist, it should.)

- data that “things” have gathered about customers will be increasingly important. The different bits gathered by individual devices will be combined to create context that makes each item more valuable than it was in isolation. Companies will be able to trade and auction this information in the way that advertisers today bid in real time for ad impressions. This, in turn, will create incentives to build devices that gather still more information. Some of this value will be shared with the consumer; some won’t.

- as devices become more important, consumers and companies will both be incented to ensure they’re always present. In fact, the most concrete idea to come out of all this thinking is a ring (the kind you wear on your finger) that buzzes when you move more than, say, 50 feet from your cell phone, thereby alerting you to the fact that you left it behind before you notice it’s missing. That’s much better than “find my phone” services that let you know it’s at the restaurant you left two hours ago. Half the people I’ve suggested this to want to buy one now; most of the others were horrified at being tracked. (The outlier was an engineer who told me it would consume too much power. Sheesh.)  This is another product somebody should build if they haven't already.  You're welcome.

If you’re a marketer, most of this probably sounds pretty appealing, albeit in a daunting, I-already-have-too-much-work sort of way. The privacy implication may bother you a bit, but most of us have pretty much given up on that front or at least walled up those concerns in a little corner of our mind that we just visit occasionally. What you haven’t seen is how this leads to machines taking over the world.

But think about it. You may recall my original scenario about the self-driving car that does uber runs on the side and keeps the earnings in its own bank account. Maybe that sounds silly, but some people might actually like their car to pay for its own maintenance out of its own account. So it’s not all that far-fetched. Once you start there, how much control does the car have? Should it make its own investments, and, if so, might it not skim some money for its own purposes or to share with the manufacturer or software provider? Do the cars pool their resources to buy stock in a car company and take over its management, ostensibly so they can get it to build better cars? It's perfectly logical: that might be the best way to reduce long-term cost of ownership. Or maybe the cars skip the ownership part and just create a fictitious “SVP of Product Development” who sends out emails to product designers. Who would even notice that no one has actually met this person? And maybe the cars conspire with the computers in Congress to insert some interesting clauses in new laws, say relating to fuel economy or consumer safety or required maintenance schedules? Again, it’s not clear anyone would ever notice that no human was ever involved. From here it’s just your standard science fiction leap to the machines making decisions that are theoretically in humans’ best interest but actually result in the machines taking control.

In short, no matter how I try to spin this, it all gets pretty depressing pretty quickly. So I don’t think I’ll be showing this movie at the next MarTech conference…although I can’t promise some machine won’t create it without me.

LeadLiaison Helps Marketing Automation Users Break the Content Bottleneck

You may have noticed that there are many B2B marketing automation systems available. So it’s not surprising that LeadLiaison prefers to be called something else – in their case, “revenue generation software”. I’m not sure exactly why they chose that term, apart from the fact that everybody likes revenue. But they do go far enough beyond standard marketing automation to justify a different label.

In particular, LeadLiaison helps marketers create content, a critical bottleneck that is not addressed by most marketing automation systems.  The most impressive feature is an outsourced content creation service, which lets marketers build an online creative brief for a particular item and then send it to a network of writers who agree to produce it for a fixed price in a few days. Identities are hidden in both directions, so the parties can’t easily circumvent the service to work together directly in the future. But prices are either reasonable (if you’re a buyer) or ridiculously low (if, like me, you're a sometime content creator), starting around $50 for a blog post. There are several third-party networks that offer this sort of service, but I’m not aware of any other marketing automation vendor that has developed their own.

LeadLiaison takes good advantage of this feature by closely integrating the resulting content into other operations. The outsourced content can be loaded directly into the marketer’s content library with access controls based on date range, number of downloads, or whether the requestor provides an email address.  There is also an option to request an email address but then grant access even if it's not provided. Content can be linked to a social media publishing process that can release it immediately, schedule it for the future, or add it to a “buffer” of materials that are released at predefined intervals. The system warns users when the buffer inventory is dangerously low, so they have time to replenish. Content is served through short URLs to track consumption and sharing. The system also tracks consumption by individuals using cookies and by companies based on IP address.

Marketers who want to build their own content are also covered.  They get powerful tools for email, landing page, Web form, and survey creation, including templates with drag-and-drop editing for different types of components.  The system can also extract the HTML of an existing Web page, insert new content such as a Web form or survey, and deploy the modified page in place of the original. That's a big deal: it means marketers can add their content into existing Web pages and forms without recreating them from scratch.

Forms can generate an alert or take another action after they are submitted. They can also include progressive profiling rules to avoid asking people questions they have already answered. LeadLiaison is adding a marketing content map that will help planning by showing the inventory of available content by buyer type of purchase stage.

Although content creation is probably LeadLiaison's most unusual set of features, the system also does an above-average job at the standard marketing automation functions: email, multi-step workflows, behavior tracking, lead scoring, CRM integration, and analytics. To look at each of these in turn:

• emails can be sent by the vendor itself or through integrations with third parties including Constant Contact, MailChimp, SendGrid, and others. Salespeople can also send through Gmail, Salesforce.com, or Microsoft Dynamics CRM. Users can preview how their emails would look in different clients. 

• workflows support multiple steps, event-based triggers, wait stages, and a wide variety of actions including lead scoring, lead distribution, list management, alerts, and calls to external Web hooks. 

• behavior tracking captures email responses, Web site visits, form completions, downloads, and video viewing through Wistia. Users can define “buy signals” based on combinations of behaviors, which in turn can be trigger actions in workflows.

• lead scoring assigns separate scores for “fit” against a target buyer profile (which LeadLiaison calls “grading”), for recency, total activity, buying signals, and specific actions the user has assigned points. Users can prioritize leads by combining these elements in a single aggregate score according to user-assigned weights.  

• CRM integration includes native connectors for all editions of Salesforce.com (not just the Professional edition, as with most marketing automation products), soon to be supplemented by Microsoft Dynamics and Sugar CRM. A Zapier connector supports integration with other systems. Salespeople can receive alerts, hot lead reports, and detailed information about Web site visitors. The system can use IP address to identify the company of anonymous visitors and will look up possible contact names at those firms from external sources including Data.com and LinkedIn. There is also an integrated phone dialer.

• analytics tracks content usage, conversions, lead distribution, email results, Web visitors on internal and external pages, and return on investment. Enhancements for more advanced reporting are also planned for 2015.

In short, this is a very mature marketing automation system for a company that launched in 2013. My take is that they learned from the experience of older products. Pricing starts at $500 per month, up to 5,000, which makes the system affordable for small companies even though the features are robust enough for the mid-market and perhaps higher. A stand-alone visitor tracking product starts at $200 per month.

Do Self-Driving Cars Pick Their Own Gas Stations?

I had a delightful and well-lubricated dinner this week with Scott Brinker of @chiefmartec fame, ostensibly to discuss the next edition of the MarTech Conference but mostly just to chat about the industry and what comes next.

Scott wasn’t too impressed by my notion of an advertising-supported toaster (see my last blog post), even though I pointed out you could segment the messages based on the type of bread the person was eating. On the other hand, I was very intrigued by his notion of marketing through services such as alerting a driver when they need gas and where to find the most suitable gas station.

Where that example got interesting was when we added self-driving cars to the mix: why couldn’t the car take itself out for gas when the driver isn’t using it, or indeed, take itself on other chores like state inspections, oil changes, and scheduled maintenance? And if it does that, how will it pick the supplier?   Sure, the owner could specify in advance, but won’t at least some owners want the car to find the best price on gasoline or respond to special offers such as coupons?

If you do give the car some discretion, how do you know it won’t make choices based on its own preferences?  Perhaps it will favor the gas station that wipes its windshield or gives it a free tire rotation, which you have to suspect feels mighty good to an auto. Indeed, how do you know your car isn't taking Uber jobs on the side, or drag racing with its car friends from the other side of town who you never really liked?  Could the manufacturer have some involvement in this, pocketing that Uber revenue or biasing those purchase decisions in return for payment from suppliers? More generally, when devices become autonomous, do marketers still address their owners or are there ways to sell to things themselves? 

There’s at least a bad science fiction story in all this (“Do self-driving cars pick their own gas stations?” with apologies to Philip K. Dick)., which I naturally proposed to Scott as a short video for the next MarTech conference.  He didn't exactly leap at the chance.

But there are also more serious issues and opportunities to consider. Perhaps interruptive marketing really will be replaced by embedded services and subscriptions which will make product selection and purchase timing decisions without the owners being involved. In some ways, it already happens: think about the choices that a doctor makes when selecting your treatments or building contractor makes when constructing your house. We already know there is plenty of trade advertising to affect those choices. As more decisions get delegated to automated agents, this may be an area we can learn from. But of course, it won’t be exactly the same, so there will be plenty of new approaches to pioneer as well.

This is definitely the kind of discussion to have over drinks. I can’t go into details but rest assured that Scott’s plans for the next MarTech conference do take this into account.

BlueConic User-Driven Marketing Maturity Model: Surprises on the Road to Customer-Centric Marketing

I’m as fond of hearing my voice as most consultants, which is very fond indeed. But the best part of my recent presentation with BlueConic was listening to the voice of someone else’s experience: in this case, the experience of more than 60 BlueConic clients, distilled into a maturity model that traced the stages they passed through on their way to full customer-centric marketing. (Click here to see the Webinar and download the related paper.)

The good thing about hearing from someone else is you find out things you didn’t already know. In this case, I was certainly familiar with the general notion of a maturity model, as a sequence of increasingly-sophisticated stages that companies pass through on their way to the highest level. And, for what BlueConic calls “user-driven marketing”, I already knew that the final stage would be a central database and decision engine that gather data from all channels and select the treatments that each channel delivers. So it wasn’t too hard to imagine that the preceding stages would start with totally disconnected channels and move slowly to complete integration. But there were still some new insights from BlueConic’s hands-on experience. Some that particularly struck me are:

• Listening first. The very first stage of the model, Level 0, involves no differentiation at all: every customer is treated the same; in fact, customers may not even be identified. BlueConic gets involved at Level 1, where treatments are tailored to the individual but each interaction managed independently within each channel. At that stage, all the central marketing system can do is “listen” to customer activities and make the data it assembles available to the channel systems to help guide their own decisions. I would have expected the central system to actually drive decisions at that stage, but BlueConic's experience is different.

• Coordination later. Level 2 of BlueConic’s model still has each channel running separately, which again is a bit surprising. What changes at this level is that  interactions within each channel are now coordinated by the central engine. It’s only at Level 3 that interactions are coordinated across channels, and even then the scope is limited to online channels. On reflection, an intra-channel-only Level 2 makes sense: marketers need several new skills to design and measure multi-interaction programs, and mastering those is a big enough challenge without also adding the complexity of managing across channels.

• Segmentation. The growing importance of segmentation at successive model stages was perhaps my biggest surprise. When I think of tailoring interactions to individuals, I think of working with each individual’s data directly. Segments don’t enter into it. But, as BlueConic’s experience reminds us, practical marketing tasks like content creation, program flows, and result analysis are organized around groups of similar customers. This ensures resources are spent effectively and you have enough volume to measure results meaningfully. In fact, the segments get increasingly refined with each maturity level as behavioral data is added (Level 2), segments are adjusted in real time (Level 3), and segments include predictions and events (Level 4). Thus, the process does move closer to treating each individual differently, but always in a segment-based framework.

• Complexity of data. This was less a surprise than an observation. Part of the presentation was a set of examples presented by BlueConic CMO Dan Gilmartin. By the time we got to Level 4, where interactions are being coordinated across all brands as well as all interactions in all online and offline channels, the example was offering a soccer jersey as a holiday gift idea to a mom reading a lifestyle Web site. Superficially, this seems like a simple, obvious thing to do.  But, on reflection, it’s amazingly complex. It requires not just knowing who the viewer is, but who she’s related to (child or spouse), the interests of that related person (soccer), and the temporal context (holiday gift buying season). That is some pretty fancy data management.

Not everything in the model surprised me. In particular, BlueConic’s experience confirmed the importance of process and organizational change to support the new technologies. BlueConic reported a steady expansion of the scope of measurements from tracking response to independent interactions (Level 1) to tracking movement through the customer journey (Levels 2 and 3) to measuring the incremental impact of each interaction on customer lifetime value (Level 4). Similarly, it showed a shift in management perspective from optimizing results for individual interactions (Level 1) to each channel (Level 2) to maximizing value for the organization as a whole (Levels 3 and 4). And, finally, it reflected a shift in control from channel managers operating more or less independently to central managers who focus on customers and segments. This all ties back to the central notion of the maturity model: that technology, process, and organization must all be aligned at each stage for the business to execute effectively.

By all means, download the Webinar and white paper, which contain plenty of insights beyond those I've just described.  Incidentally, if you're wondering about that interactive toaster, I was already aware that you could get static custom images on bread and have since discovered that there are some higher tech options.  I see no technical reason one of these couldn't be connected to the Internet to deliver dynamic messages sent by an advertiser, significant others, or favorite government agency. 

Radius Provides High Quality Data on Small Businesses

When I first spoke with Radius just over one year ago, the company had already pivoted from its initial concept as a mobile app to connect consumers with local business events, to building a comprehensive list of small businesses and their attributes. Fast-forward twelve months and the company has again adjusted its offering, now presenting itself as a “marketing intelligence platform” that helps business marketers find prospects who are similar to their current buyers. This latest vision was appealing enough to attract $54.7 million in funding in September, bringing the announced total to over $80 million. So I’m guessing Radius will stick with this approach for a while.

What Radius does will sound broadly familiar to loyal readers of this blog: it scans social media, Web pages, government records, and other online sources to build a list of more than 20 million U.S. businesses and their attributes. It supplements these with conventional data sources to capture businesses with a limited digital profile. In its current incarnation, Radius also imports a list of won and lost deals from each client’s CRM system (direct connection to Salesforce.com, batch imports from others) and shows how well each attribute correlates with success.

Users can review the attribute list, create segments based on attributes, and analyze the attributes of each segment as a group. They can also flag existing segment members within the client’s current CRM database and import segment members who are not already in the client’s CRM (a.k.a. “net new prospects”). The imported records include basic company information and other attributes the client has preselected, but the system will not correct or enhance existing CRM records. Segment membership is adjusted automatically as Radius updates its data, which happens weekly. The system does not store fixed lists of segment members at a point in time, although users could achieve this by tagging records in CRM as they are imported. 

And that’s pretty much it. No list of the most important attributes, no predictive modeling, one contact name per company, no alerts based on buying signals, no campaign analysis: just company information compared to your own customers, an way to build segments, and an option to purchase new prospects. The company plans to address some of these gaps but has not released the details.

Whatever its limits, Radius has attracted some big-name customers, most notably American Express, as well as all that funding. The primary reason seems to be data quality: the company says it can usually match 80% to 90% of the businesses in a well-maintained CRM system and that client tests have shown it is more accurate than competitors. This is both impressive and important, especially where small businesses are concerned. Available data includes basics (address, phone, industry, company size, revenue, contact name), Web activity (presence of a Web site, Facebook and Twitter accounts, use of daily deals and check ins, and average review ratings), and technologies used.

The system has some other advantages.  New clients are deployed in 24 hours, including the time to import CRM data and calculate success rates by attribute. The user interface is attractive and intuitive.  Pricing starts at $15,000 per year for small enterprises.  It is based on the number of company records in the client database, so it doesn’t increase based on how heavily the system is used.  It also includes use of the system software and credits for some number of new prospects imported from the Radius database.

In short, Radius strikes me as a solid solution for what it does, which is provide targeted company-level prospect lists and profiles of your current customer base. If that’s what you want, take a closer look. If you want to know more about trigger events or individual contacts or want lead scoring or other types of predictive modeling, you’ll probably be happier with something else.

Sailthru Offers End-to-End Omnichannel Personalization for B2C Marketers

I know this is blasphemy, but I’m beginning to have doubts about solution selling – the idea that marketers should describe the customer problems they solve, not the features of their products. The issue, at least in marketing technology, is that all systems address pretty much the same general problem of sending the right messages to the right customers (in the right time, place, medium, device, language, tone, etc.). This means that solution statements sound pretty much alike, even when the actual products are different. It’s left up to the poor buyer to figure out what each product does and whether that is something she truly needs.

Sailthru is a good example. The corporate home page says “Sailthru makes it easy to personalize every channel for every customer,” which is accurate enough.  But plenty of other companies also help with omni-channel personalization.   A marketer looking for personalization solutions might add Sailthru to her list of options, but wouldn’t know whether it's a good or poor fit without digging much deeper.

On the other hand, resolving that sort of ambiguity is what keeps me in business.  So perhaps I shouldn't complain.  In any event, there are several differentiators that determine whether a product like Sailthru is suitable for a particular situation.

• customer profiles. Sailthru builds a history of information about individual customers.  You  might think that would be done by all personalization systems but it's possible to do something that can reasonably be called “personalization” using only anonymous information such as traffic source, search terms, location, or Web pages viewed during a visit. Sailthru goes beyond this to store behaviors over time.  These are linked across channels to a customer identity that is usually known at the start of an interaction. The identity might be available because the customer is interacting with a mobile app for which she has registered, is responding to an email or text message that was already tied to her identity, has logged into an ecommerce Web site, or is known through a cookie that was previously linked to her identity. Sailthru generally does not deal with anonymous customers. It can store several identifiers for the same customer, which is how it coordinates interactions in different channels. The identifiers would be linked through “hard” matches such as an email address provided when registering a mobile app or an ID number embedded in a Web link in an email.  "Fuzzy" matching, which attempts to link identifiers that have not been directly connected, is generally avoided by Sailthru.

• data store. Sailthru stores data in MongoDB, a “No SQL” database that can handle nearly any data type and can easily add new “fields” (not really the correct term) without formally defining them in advance. This makes it extremely flexible, which is very important in the fluid world of marketing information. Mongo is also fast and scalable and good for analytical processing in general. A fair number of multi-channel personalization systems use Mongo or something similar, but many others use conventional relational databases (which are less flexible) or other data stores.

• data sources. Sailthru gathers most of its data from its own tags placed within emails, Web pages, and mobile apps. This distinguishes it from systems that rely primarily on feeds from external systems via API connectors or batch files. Technical users can still set up a feed using API calls or JSON posts when necessary. Prebuilt integrations are available for Magento ecommerce and WordPress, with others on the way. There are no standard integrations for marketing automation or CRM. The system usually stores emails sent, Web pages visited, purchases, content read, and mobile interactions. The system can scan, classify and tag company’s marketing contents and then use the tags to build a customer’s content consumption profile. It can do similar tracking based on merchandise category tags from ecommerce systems. Users can also set up custom variables derived from original inputs.

• predictions. Sailthru has a recommendation engine that uses customer history to suggest the product or content they are most likely to select next. It has recently released the beta version of a predictive platform that automatically generates probability scores, rankings, and estimated values for nine actions such as making a purchase within the next 24 hours, opting out of future contacts within the next week, and expected revenue within the next thirty days. These can be used in segmentation and message selection. General release of the prediction tool is planned for early 2015. Predictive models are rebuilt and records are rescored nightly, with no changes during the day in response to new customer activity. Recommendations do adjust in real time to customer behaviors. The system can create control groups to measure the impact of recommendations on long-term customer behavior. These predictive capabilities are among the biggest differentiators for Sailthru: predictions and recommendations are oriented to consumer marketing, not B2B lead scoring, and Sailthru doesn’t (yet) allow clients to choose what they wish to predict. On the other hand, the modeling is fully automated, while many other systems require at least some manual set-up for each new model.

• message selection. Users can define lists based on any data in the system and then send a specified email or mobile message to each list. They can also export lists for Facebook promotions or to other channels.  Messages and Web pages can contain real-time recommendations and can also adjust their contents based on data and scripts written in Sailthru’s own Zephyr language.  Marketers should look closely at this aspect of Sailthru: while powerful, it's not creating rules to send different content to different segments, doesn’t send sequences of messages over time, and doesn’t support real-time interaction flows.*  Be sure you're getting the personalization features you need.

• message delivery. Sailthru builds and delivers email, mobile, and Web messages directly, rather than sending lists or recommendations to other systems. Many marketers will like this,  since it avoids the need to integrate with another product. But marketers who have want to use other delivery platforms may not be happy.  This is one reason I haven’t classified Sailthru as a customer data platform: although Sailthru does a great job of building a unified customer database, most CDPs are specifically designed to work with other systems when sending messages.

• data access. Sailthru lets clients export lists based on profile data, can display individual customer profiles, and provides some limited API access to the profiles. But it doesn’t support mass exports of the profiles or allow external queries of the profile database. This is the other and more important reason I don’t consider Sailthru a CDP: making the database available to external systems is the very core of the CDP concept.

• pricing and company background. Sailthru was founded in 2008. It currently has about 400 clients, mostly in ecommerce and media. Pricing is based on the number of active profiles and (unlike many personalization products) does not increase as clients support more channels. Prices begin around $30,000 per year. 

___________________________________________________________________________
* More precisely, it can't do those things within a single campaign flow, or what Sailthru calls a "smart strategy".  Each strategy identifies a single event which triggers a single action, such as sending an email.  Sending different emails to different customer segments would require setting up a separate strategy for each segment, each linked to its own message.

**  Similarly, a sequence of messages would require setting up a separate strategy for each step.  In both cases, it would be up to the user to write event conditions that ensure the right people are selected each time.  Real time interactions would also need to be created by defining criteria that link a series of unconnected events.

**Although users could use the Zephr scripting language to build a single message that delivers different versions to different people.  This is functionally similar to delivering different messages but is harder to manage since the variations are buried within the script.