Wednesday, May 25, 2016

CRM Evolution Conference: Mobile Really Does Change Everything About Marketing

I snuck down to Washington DC yesterday for a few hours at the CRM Evolution conference, where a critical mass of industry experts triggered a chain reaction of interesting thoughts. 


The first was that customer systems should read most data directly from the system that created it rather than loading that data into a master database. This isn’t really a new idea – it’s called federated access and has been around for decades.  But I’ve always considered it problematic because source systems might not be easily accessible and source system owners often worry that direct external access would slow their systems’ performance. Moreover, operational source systems often don’t keep old versions of important data that changes over time (such as lead scores or contract expiration dates), making historical analysis difficult if that data isn’t stored elsewhere. Despite these issues, several practitioners and vendors at the conference said they were using the approach and had found it more practical than moving all customer data into a central repository.

I’ll guess that more open system designs and higher performance technology have made direct access to source systems more practical than it used to be.  It’s certainly true that the sheer volume of customer-related data has increased to the point where replicating it all into a central system would be a massive project. Indeed, I’ve been telling clients for some time now that they will need a mix of consolidated and federated sources, with federation clearly the right choice for contextual information that is only relevant in a small number of situations. For example, you wouldn’t store the minute-by-minute history of weather in every location if it were only relevant at times and places of customer interactions. Instead, you’d look up the weather in the customer’s location when an interaction began and store it as part of the interaction history. It’s true you might miss some interesting patterns – perhaps raincoat sales spike the weekend after a big storm, which you wouldn’t know if you hadn’t tracked weather during the preceding week. But such insights are probably uncommon and there would be other ways to find most of those patterns without storing massive quantities of largely-irrelevant detail.

But the argument I heard this week was stronger than that.  It was that even information core customer information such as purchases should be referenced rather than copied. The ultimate expression of this would be a central customer record that only stores the identifiers needed to find customer data in external systems. I heard at least one vendor say her system worked this way and it's just fine, although I suspect it may have a little more central storage than she described. Other people took a more moderate approach, stating they will copy data into a central system but only if there's specific use for it. But treating replication as an exception is still a reversal from the traditional approach of treating replication as the default.  In practical terms, it means marketers need to look more closely at the federated access capabilities of systems they consider and at how those systems will deal with history data and cross-channel identity matching, which often relies heavily on historical information. So a bit of attitude adjustment may be in order.

A more profound (or, at least, less technical) chain of thought started with an “influencer panel” observation that mobile devices are now the standard tool for doing everything.  That doesn’t sound too controversial until you realize it means that mobile is no longer a “channel” or a “trend”.  Instead, mobile is simply how things work whether the interaction is on the Web, in email, in social media, by text message, or, for the Luddites among us, by voice.

This matters because mobile interactions are inherently different.  The small screen means they must be simple and on-the-go use means they must be quick. Further reinforcing these trends is the customers’ increasing expectation for personalization.  This expectation also means that mobile (and, implicitly, all other) interactions must exactly match the customer’s needs of the moment.

Put these together and you come with a goal that might be called “precision”: interaction designs and interfaces that give the customer what they want and nothing else. Imagine a painting that’s covered with a sheet of paper with one tiny hole cut out.  The paper hides most of big picture but lets the user see the one detail she cares about at the moment.  She can also move the paper to see different details at different times. The customer’s experience is made simple and direct – at its best, she sees one choice (the thing she really wants or needs) and a button to accept it. Yet the picture behind the paper can be immensely complex.

This vision has (at least) two implications. The more obvious is that it takes incredibly powerful technology to anticipate the customer’s needs.  This is where things like context and machine intelligence will come into play. The location- and situation-aware nature of mobile technology, especially when it’s connected with other devices through Internet of Things, will provide the information needed to understand the customer’s precise situation. Machine intelligence will provide the processing power to interpret this data correctly, continuously, and for millions of customers at a time.

As I said, that implication is important but it's not exactly news. The second implication has been less discussed.  It's that when you strip away everything except what meets the customer’s present need, you lose the opportunity to communicate other messages that might serve your long term purposes. Metaphorically, the hole in that piece of paper is so small that the customer sees only one button, and not whatever advertising might have previously surrounded it. So there’s no opportunity for branding or nurturing or educating the customer – and, perhaps most frightening for a marketer – no way at all to reach potential new customers.

It's as if the Mona Lisa were presented in personalized parts – with geologists shown only the mountains, hairdressers shown only her hair, ophthalmologists shown only her eyes, and plastic surgeons shown only her nose.  Each might come away satisfied with their Mona Lisa Experience, and perhaps even delighted. But I think we'd all agree that something would still be lost.*


Conversely – and this is a third implication – those narrow interactions provide less information about the customer. (The marketer is looking back at the customer through that same small hole in the paper.)  This makes personalization even harder. 

Maybe you think I'm overreacting.  After all, operational interactions where the customer has specific goal within an established relationship are not the only thing people do. But think how time-starved most people are today and how little attention they have for anything beyond their immediate agenda. “Interruptive” messages like display advertising and most marketing emails are already easy to ignore.  They'll be even easier to avoid as screens get smaller and automated assistants get better at screening out things their masters don’t want to see. And even when people are purposely searching for new information, they will rely on ever-smarter systems to many of the preliminary choices.  The days of buyers leisurely gathering a wide variety of information, slowly forming opinions about their options, and interacting with your marketing materials and sales people along the way are already gone. The buyer’s journey isn’t a stroll through the garden smelling the flowers and picking whatever fruit looks ripe: it’s a dash to the store pick-up counter where she grabs a package that someone else has already assembled. If there’s any good news here at all, it’s that journey mapping just got very, very easy.

I’ve covered some of this territory before in my discussions of trust-based relationships and marketing to machines. But even before we get to the point where humans are completely cut out of the buying process,  we'll have the problem of how to optimize the customer journey. The trick will be to deliver value during every interaction – to optimize the journey from the customer’s perspective, not the company’s.

Customers who haven’t bought yet (okay, they’re really prospects) still have a specific intention when interacting with us – presumably to learn something about our company or products so they can decide whether they want to do business.  They’ll presumably be a little more understanding than current customers if we can’t guess exactly what they want, but they’ll still have high expectations and little patience. So marketers and their systems will need to gather as much information as possible, both directly and from external sources such as intent data.**  And we’ll need to use every scrap of information as fully as possible to deliver as much value as we can.

Specifically, we want to keep prospects engaged so that each offer they accept leads to another offer they also accept.  Beyond building our own relationship, this will consume their limited time so they can't use it to research competitors. This leads towards materials like interactive content (which both is engaging and gathers information to tailor the next offer) and metrics like engagement.

In this world, the traditional view of the buyer’s journey as a sequence of steps is almost wholly irrelevant.  Our job as marketers is to meet the customer’s needs in whatever sequence she presents them, not to push her down a predetermined path. The measure of success is the ability to keep someone engaged – following the simplistic but (I think) irrefutable logic that prospects who stop being engaged never become customers.  It would be easy to base an optimization methodology on this approach.

Of course, a goal beyond avoiding disengagement would be encouraging purchase.  In addition to meeting the customer's needs with each interaction, we want to shape the evolution of those needs in the customer’s mind, so at some point her "need" will be buy our product. This provides another, more conventional metric to guide optimization.  But even the purchase need, and the resulting interaction, is just one step among many: marketing in this world is seamlessly integrated with the rest of the customer experience – and all customer experiences are part of marketing.

_____________________________________________________________________________

*Since marketing technologists have their own tunnel vision, my next thought is finding a technical solution, perhaps presenting a thumbnail of the full image linked to underlying videos.  What's scary is this approach, which I intended as joke, has probably been applied by museum curators as a serious solution.

**Which happened to be my topic at CRM Evolution. You can download my "Understanding Intent Data" slides here.)


No comments: