Hitmetrix - User behavior analytics & recording

5 Top Marketing Technology Challenges

Not every marketing technology challenge needs to be seen as a marketing technology problem.  

Some are better described as opportunities, even if there’s room for improvement in the technologies. Marketing automation opens up the possibility of creating, testing and optimizing customer-facing content, from simple messages to complex campaigns, on a vast scale. Programmatic holds out the promise of personalizing those messages and campaigns at high speed, at the right cost for brand and publisher.

Some challenges, on the other hand, do look like headaches. And they’re enduring headaches.  Yes, there are palliatives, and yes, plenty of smart vendors are working on promising solutions–but they’re going to be with us for a while.

I made my own list of the top 5 challenges of the latter kind, and then measured it against some benchmarks by looking at an extensive new report and asking around.

1. Getting a single view of the customer

Call it profiling, call it identity modeling, call it identity linkage–the problem is that the pesky customer won’t keep still in one place, even when we make them give up personally identifying information in return for registering or logging in. The holy grail of digital marketing–providing a unique and relevant experience for each individual customer–remains tantalizingly out of reach as long as the customer’s identity remains fragmented across devices, platforms, browsers, and apps–not to mention physical touch-points like phone calls and store visits.

99 percent of companies believe that a single customer view is important to their business, while 24 percent claim to have a single view today

In its 2015 Digital Marketer report, marketing suite vendor Experian identified the lack of a single customer view as the top barrier to cross-channel marketing. Based on a survey of more than 1,000 marketers worldwide, Experian found that 99 percent of companies believe that a single customer view is important to their business, while 24 percent claim to have a single view today.

Given the obstacles to linkage, especially across mobile channels, I wonder if that’s an optimistic 24 percent.

Customers–and, indeed, prospective customers–switch devices and browsers at will; use Twitter, Facebook and other logins interchangeably; often browse anonymously; and can disappear at any moment from being present on the web into the black hole of a mobile app.

At a recent presentation by the mobile identification vendor Parrable, I heard an audience member say that he didn’t doubt companies could track him effectively online. “You think a roomful of scientists couldn’t figure it out?” Well, maybe the NSA is following in our footsteps, but of course the automation of personalized marketing assumes being able to take a single view of thousands, or hundreds of thousands of customers, preferably in real time. Programmatic means being able to optimize and send messages to the right customer, in the right place, at the right time, across large audiences.

And of course the fragmentary in-house view of the customers across silos, from call centers, through sales teams, to CRM systems, only adds to the problem.

Increasing probabilistic success in modeling consumer profiles across platforms and/or browsers/devices is the future

Jerry Jao, CEO of Retention Science, deals for the most part not with acquiring but retaining customers. He told me that: “Increasing probabilistic success in modeling consumer profiles across platforms and/or browsers/devices is the future. Every business is trying to understand their customers better so they can assess customer lifetime value, understand the best way to engage them, and ultimately, create more loyal customers. It’s incredibly difficult to get all data into one place–simply because most of the data sits on all different types of platforms, and are collected differently and in different format. It’s very costly and time consuming to consolidate data and make them speak to each other. So this is key, and why it’s challenging.”

Bertrand Hazard, VP of Marketing at Trust Radius, which aggregates real user reviews of business software products, agreed: “Tracking and targeting consumers and buyers across channels and devices is definitely a challenge. A lot of vendors are trying to tackle this problem from different angles. Some like Adobe advocate a suite based approach using their own products, others emphasize integration.”

Of course, automated solutions for modeling, reconciling, and authenticating identities do exist (Experian has one). But success–“probabilistic success,” as Jao says–relies on the quality of the data.

Which brings us to:

2. Getting enough good data

Big data. Anyone need to hear that phrase again? Or learn how many gigabytes of data we’ve created in the last few years compared with the whole of recorded history? Yes, about 90 percent of data ever created is relatively new. Thanks to online activity, and especially use of social channels, brands now have more customer data than they know what to do with.

What we need is less data science and more data engineering

The problem is, a lot of it stinks. First party data–the information a prospect or existing customer supplies a brand with directly, from name and email address to credit card details, plus whatever the brand can glean about on-site behavior–is invaluable. But it’s not big. On the other hand, the third party data which can be bought from vendors is increasingly reviled as–deep breath–incomplete, out-of-date, inaccurate, duplicative, irrelevant, or just mispelled. The founder of a New York-based real time analytics company told me, “What we need is less data science and more data engineering.” In other words, what Experian refers to as the “daunting” task of data hygiene.

The match rate of data is typically less than 50%

What about second party data? That would be the immensely valuable data which users happily share with Facebook, Google, Tumblr, Pinterest, and Twitter, and so on, and which those companies will only release in ways constrained by price (high) and policy (privacy agreements). And again, however good the data, mapping it onto a linked customer identity just brings us back to the first problem. Jao said: “The match rate of data is typically less than 50%. It’s harder than most people realize. This is definitely something companies are trying to do a better job on.”

On the other hand, maybe it’s a glass half full situation. I’ve heard Jeremy Hlavacek, VP of Programmatic at the Weather Company, say more than once at conferences this year that probabilistic identification can be enough. Bad data is better than no data, and identity linkage at only 10 percent accurate can be very valuable across a large sample.

Let’s say you’re confident enough that you’re tracking individuals throughout the customer journey at a reasonable rate of accuracy. Now all you have to do is optimize the content you’re delivering to create sensational customer experiences.

Wait. Where’s the content? Well most of it is in marketing folders. Some versions are being shared among team members via Drop Box and other services. Those great videos the agency commissioned last year, well, the agency has them. And there are different versions of slides, presentations and reports, as customized by individual members of the sales teams, on their laptops and tablets. Which means you need to solve the next conundrum.

3. Getting on top of your assets

It’s all very well signing up for a social relationship management solution which empowers your team to respond instantly to the global conversation. It’s all very well tracking customer behavior on webpages, inside apps (yes, that visibility is coming), and in response to emails. But if you can’t lay your hands, pretty much instantly, on the content and messages prospects need to see, and deliver them to the right place at the right time, your team’s going into battle empty-handed.

Digital asset management–DAM–is, to many people, at the very heart of content marketing; or optimizing the customer experience, if you prefer. It’s a topic we’re writing about more and more here at The Hub, because it’s evident that even a relatively small asset inventory can’t be managed on an Excel spreadsheet. For companies with large sales teams and multiple locations–and especially in sectors like health, where compliance is paramount –DAM needs to be thought through and automated. After all, there’s not even any point collecting your content all in one place if you can’t instantly find what you need; know whether it’s been approved; and know whether it’s been amended. What’s more, if you’re not on top of versions, you can’t know what customers respond to, and can’t optimize the content for future use.

But assuming you’ve identified your customers accurately, turned on a good, clean data stream, and corralled the content you need to deliver, it’s all systems go, right? Well, no, not if you want eager prospects rather than an angry mob.

4. Getting frequency right

The rather quaint concept everyone talks about is the “sweet spot.”  There’s a sweet spot for frequency of email messages. There’s a sweet spot for frequency of targeted online ads. Too few, and the message might not stick; too many, and you have an unhappy prospect. We’ve all experienced this. In the weeks after I research a product for an article, I’m acutely aware that I’m receiving targeted ads based on the (false, unfortunately) assumption that I’m a prospective purchaser. That’s fine–until I’m still being targeted weeks or months later.

Customers are everywhere, but don’t want to be reached everywhere 

The frequency question is, of course, just a part of the online privacy question. “Customers are everywhere, but don’t want to be reached everywhere,” said Jim Caruso, VP of Varick Media Management, a programmatic vendor, at a recent conference. Privacy, however, is by no means top-of-mind for hard-pressed marketers.

“Being sensitive to privacy and frequency of contact concerns, without losing marketing momentum? I feel this was a bigger concern last year and the year before,”Jao told me. “Nowadays, I hear people worrying about securely collecting the data, but less about privacy concerns… Most marketers understand that data is the only way leading to better marketing so they’re investing in the time and money to get more data, versus being sensitive about customer privacy.”

My prediction: if programmatic really takes off as a way of delivering personalized customer experiences at high speed and high volume, brands which take the opportunity to firehose prospects will get a bloody nose–no matter how relevant their messaging is. Sometimes people just want to be left alone.

But let’s assume the savvy market has a sensitive handle on frequency of touch too. How is he or she doing? Who knows?

5. Getting a fix on ROI   

The question about what a brand is getting for its marketing dollars predates marketing technology by, oh, two hundred years or more. 

Nevertheless, speaking with digital marketers, I’ve repeatedly been surprised at the cavalier attitude to analyzing how their efforts are driving ROI. I spoke with a marketer for a brand with outlets and agents nationwide. His creative team was first class. All his engagement indicators were healthy. Customer acquisition was satisfactory. Did he know whether what he perceived as success was being driven by his cross-channel digital engagement, and how much by the traditional calls to action in which the company was investing? They hadn’t really looked at that.

If you have the technology to address the first four challenges above, there’s really no excuse for not being able to track at a surprisingly granular level, the ROI not just of campaigns, not just of individual messages, but of specific iterations of messages as received by individual customers within very specific contexts (at work, in the morning, on a smartphone, etc). After all, if you have no sense of whether these fine-grained touch points are driving the desired results, how are you going to continue to optimize the customer experience?

First, though, you need to be clear on what kind of return you’re looking for. For most marketers, it’s always been a simple matter of conversions or revenues. But in this age of the empowered customer, ongoing engagement through the full life cycle, from prospect to post-sales service, may be more important than a single purchase decision. An engaged customer is not just a revenue source; he or she can also be a loyalist and advocate. If you believe (as you probably should) that online peer recommendations, especially in a social context, are more effective in driving sales than purchased media, then KPIs for your content should extend beyond conversions–and they probably do: at least to measuring responses, shares, and possibly sentiment too.

KPIs for content might relate to social reputation, positive brand awareness, brand advocacy, and customer reactivation–not just to driving more product off shelves. What do you measure that matters?

The conclusion? Complexity. Complexity for the customer, and even more so for marketers, for the analysts  helping them make sense of the inexhaustible data stream, and for anyone charged with building responsiveness, efficiency, and scaleability into the marketing technology stack. 

But let us know what you think the challenges are. Is this list correct? Is it even close to complete? There’s a message board below, or email me: [email protected].

Total
0
Shares
Related Posts