by Gary Angel

The four P’s of analytics

Aug 11, 2016

The topics dominating discussion at the enterprise digital analytics table are prioritization, personalization, people and perspective.

data analytics charts
Credit: Thinkstock

Few fields change as fast as digital. New channels, new methods, new business models — and all of it demands new methods of measurement and analytics. As new technologies and practices disrupt the field, digital analytics practitioners adapt. In any given year, a few themes dominate, and right now, the topics dominating discussion at the enterprise digital analytics table are four P’s: prioritization, personalization, people and perspective.


Analysts face a supply and demand problem. Executives have an insatiable demand for analysts’ insights, but analysts have a limited capacity to supply answers to a seemingly endless stream of questions. It’s a good problem to have — I’ve been around long enough to have seen how much worse it was when nobody wanted digital data.

In almost every conversation I have with digital analytics managers, keeping up with the demand for ad hoc analysis is one of their biggest challenges. And this isn’t just about providing answers. As one of my colleagues says, 90 percent of any problem is defining the question. If you can pinpoint exactly what someone needs, then the path to finding the answer (or establishing that the required data doesn’t exist) is quicker and more efficient. A lot of the questions analysts field are like flour, sugar, chocolate and eggs. They need a little mixing and processing before they’re ready to be baked! Digging deeper into every request should be standard operating procedure for every analytics department.

Another key technique for managing ad hoc requests is transparency. Making your backlog publicly available can help executives see how their requests fit into the bigger picture. One manager I talked with recently provided a quarterly update to her leadership team to give an overview of how long-term projects are progressing and how much bandwidth there is for short-term, timely needs. I like a lot about that approach. Not only does it manage expectations, it’s a helpful reminder to leadership of how much analytics is getting done. It’s also a way to carve out space for deeper, more complex analytics projects. No matter how valuable all that ad hoc information getting is, it tends to be forgotten when budgets are being determined.

That’s why prioritization isn’t just about how to get work done. It’s a necessary step in giving analysts the time and space they need to be creative and do their best work. Force your analysts to constantly cycle between lots of questions, and not only will they provide fewer answers, they’ll never get around to tackling the big problems that matter most to the organization.


Of those big problems, the one that’s clearly at the top of the list is personalization. One analytics leader memorably remarked that, “Personalization is the big sexy in the room.” Yeah. But in the trenches, it’s surprising how many digital analytics managers I talk to are still unsure about how to get value from personalization.

There are plenty of stories about personalization projects that become unwieldy monsters with lots of effort, lots of technology and results that aren’t necessarily more impressive than what a modest A/B testing program can achieve. It seems clear that, in the spirit of much of today’s digital enterprise, smaller, more agile, more organic efforts that gradually deepen the breadth and depth of segmentation and the corresponding personalization are more likely to succeed than big bang efforts. On the plus side, when personalization works, it works. We measure many sites, and in almost every case where a digital property significantly improves its performance, some form of segmentation and personalization is the driver.

Beware, though, of falling prey to the “me-too” effect when thinking about personalization. Everyone is so familiar with the product recommendation strategies (“people who bought/viewed this item also bought/viewed this”) on giant internet sites that people tend to think of that type of basket-based recommendation engine as synonymous with personalization and not just as an example of one style of personalization. For enterprises without large product sets, thinking of personalization in terms of a product recommendation engine can make it hard to see how personalization is relevant. Every experience can be varied and made more personally relevant — not just product selection. I think there’s irony in viewing personalization with a one-size-fits-all lens!


If personalization is the thing we’re mostly struggling to do, the biggest concern I hear about getting personalization done is finding the people to do it. I don’t talk to many digital managers who aren’t having a hard time finding (and keeping!) talent. As with any white-hot industry, the issue of human capital is “what keeps us awake at night.”

When you have a mismatch between demand and supply, price goes up. That’s obviously a good thing for us practitioners, but it puts a strain on managers seeking to hire and retain talent — especially when they find themselves restricted by corporate HR policies that don’t appropriately match job titles to market salaries and don’t create reasonable career paths for in-demand analysts and data scientists.

So if you can’t buy them, the thinking often turns to making them and, once made, keeping them. When it comes to training analysts, I hear an almost overwhelming consensus that résumé skills are terrible predictors of analytic chops. Teams that consider themselves successful in hiring and training analysts almost all seem to focus on finding curious, logical thinkers — not on any particular degree or academic achievement.

Of course there’s nothing worse than finding great young analysts, honing their skills, and then having them walk out the door just when they are starting to really contribute. But it’s mostly not salary that drives attrition. The number one source of job satisfaction (and retention) that analysts cite when talking about why they stay or leave is whether the work they do makes an impact on the business. Talk about a vicious circle! If your organization isn’t great at using and operationalizing analytics, you’re likely to lose the very people who might make a difference.


Another topic that keeps coming up is voice of the customer (VoC) and the increasing challenges organizations face in hearing their customers accurately. Most organizations have been doing online intercept surveys for a while now. But changes in VoC practice are afoot.

It’s becoming standard practice to realize that site intercept surveys are just one part of getting the VoC right. Call center, social media, usability testing, mobile and employee feedback are increasingly getting noticed, attended to and integrated into better and broader views of customer thinking. Still not fully understood, in my opinion, is how poor traditional survey techniques are as sampling instruments compared with online sampling. When you survey, getting a representative sample is everything. And while problems with online sampling are real, it’s become nearly impossible to get decent samples with traditional phone or physical intercept methods. Watching the pollsters struggle to get predictions even remotely close to right in the 2016 election should get people to reflect on the likely accuracy of those offline survey instruments they’re still relying on.

Right alongside the trend toward more VoC sources is an increasing recognition that there is more to customer attitudes than how they rate you. VoC is the most powerful tool in your kit for shaping big, strategic decisions around product, marketing and servicing. If you get over-focused on a single performance metric, you’re largely missing the point.

You might think the increasing focus on big data has relegated traditional VoC to the scrap heap. Not a chance. You can’t always infer what people think (and what made them decide) from what they did. Advanced analytics and big data champions are often the strongest proponents of having a deep understanding of VoC.

Staying focused on the customer perspective. Using that knowledge to make experiences more relevant and more personal. Keeping your analysts productive by managing ad hoc analytics and just flat out keeping your analysts. That’s what’s top of mind in digital analytics right now.

Unless, of course, it’s omni-channel analytics, handling a multi-device world or making those mobile devices really perform for your business… well, that’s what’s great about digital and digital analytics.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.