Marketing Analytics Roadmap: The Art of Data Interpretation

Marketing analytics refers to the daily and monthly minutae of a never-ending and ever-changing flow of human behavioral data. We can now measure everything. In spite of this ability to collect data, marketing analytics is not an exact science, just like neuroscience is not yet exact. Of course, we now know a lot more about the brain but it is still a bit of a guessing game when it comes time for treatment. And treatment is the goal – the whole point of analytics is to improve performance, whether that means email opt-ins, leads or sales.

A brief nondigital marketing analytics example: If we move a line of cans from the grocer’s first shelf to the third shelf in the canned foods aisle – the third shelf being the average elbow height for an American woman – and we sell 350 more cans that week, it does not take a degree in statistics to conclude that is where we want our cans to be stocked.

The same basic premise applies in the digital marketplace. Marketing analytics helps us do it better by testing and tracking performance. It is the new big-deal position in marketing departments. Lots of new hires, consulting services, and scrambling for knowledge is taking place. There is a lot of money being spent on marketing analytics right now and for good reason: Marketing analytics has become a make-or-break skill set in every competitive market.

Marketing Analytics data is easily tracked (and free in most cases) but not so easily interpreted. Interpretation depends more on the end goal of the client than fancy mathematics. This is one common disconnect in the analytics roadmap: Clients often do not know what the optimal end goal should be, what their margins will bare on a per lead basis, and what best qualifies a worthwhile lead. This is not a slight on clients but the frustrating nature of a measured approach. For our purposes here, we will assume our end goal is always either data capture or sales growth.

The difference between diagnostics and improved performance is not unique to marketing. My friend Chris Finlayson, a former Cornell Professor and Ocean Scientist (“and Mainer” he would add) has the same obsession in his motorcycle shop – Chris owns the best vintage japanese motorcycle repair shop near Asheville, NC. About working on vintage motorcycles, Chris points out that “questions arise every day – they are hard, sharp, unambiguous, and unavoidable.” The motorcycle either starts or it doesn’t. The engine runs smoothly or it makes a worrisome sound. Once we have taken the proper diagnostic tools out of the tool box, it’s mostly a question of intuition, trial and error (we call it “testing” or “split-testing” in marketing analytics), and maintaining focus on the desired result. I caution any organization looking to employ analytics not to get caught up in the tools of the trade because the benefit is derived from a deft hand, methodical thinking, and the creative content on the page. Focus on the goal, not the tools.

Before we get to the fun stuff, here are the main reasons most companies want to hire a marketing analytics professional:

1. They have read the article in The New York Times
2. They believe they could be doing better, “ya know, sales-wise”
3. They want to maximize their marketing dollars, improve ROI

We are all hoping to be the proud owner of a well-oiled, fast-moving marketing machine that costs the minimal amount of money to maintain while driving up the top line (sales). First things first, of course, you have to get started with an expert or team of experts in the following fields: PPC, User Experience, Copywriting, Marketing Strategy (not the hypothetical MBA version but the entrepreneurial “sales funnel” variety), Split-Testing, SEO, Keyword Research and basic programming knowledge, and then finally comes the marketing analytics stuff.

I see many marketing analytics job listings looking for statisticians but I am still not convinced this requires a doctorate. Impressive as they are, most mathematical experts that I have crossed paths with are not exactly wired for this type of work. It is the direct marketing set that tends to be the most apt here, at least that’s where many of the field’s best practices were developed. Nothing against PhDs or MBAs but I would also rather have Chris adjusting my 30 year old brake caliper than a tenacious young person fresh out of the local motorcycle mechanics training course. I do, however, believe there is a place for PhDs and MBAs in the more predictive end of the field.

The predictive end of the field is all about Modeling. If you are trying to use modeling to predict P&L, set performance expectations, et cetera then you do want to hire for an academic approach.  This type of need crosses beyond the realm of marketing analytics and into the realm of business analytics. The need for such deep understanding is uncommon in most marketing departments. More often than not, marketers should focus on driving demand, gaining consumer insight, and applying deeper localization and segmentation to their campaigns and offers. Conversion optimization is more of a consumer-focused discipline than a data-driven one.

The marketing analytics stuff

The marketing analytics stuff tracks traffic patterns (clicks) through to conversions (the goal: social media engagement, email opt-in rates, sales). The tools track the behavior of your visitors and those who’ve seen your ads. With the tools in place, we then get down to improving the performance of our ad campaigns, landing pages, and sales funnels. It is a very different skill to read a traffic flow chart than to plan or refine one. Some designers can be trusted to get it right and some cannot. Some web developers/programmers get it and some don’t. You want somebody that can do both – read data and initiate improvements.

Most companies would not trust a newly-minted mid-level hire, nevermind a consultant, to redesign their company’s homepage but that’s what we’re talking about here much of the time. My newest client and I are well into an initial split test of their homepage, now three months into the client engagement, and we are seeing conclusive results (a large, multiple-of-ten percentage points, increase in data capture conversions).

I hear you, “Nevermind the homepage, Aron. What else do you have in the way of marketing analytics? I want leads and sales, Aron.”

Traffic Flow, Lead Conversion and Acquisition

Where is the traffic being funneled? Where is it coming from? Is it paid traffic from advertising or free traffic that comes in through search engines and other referrals? It matters because different messages will likely appeal to different traffic sources. When considering paid traffic, we must consider our keyword research which tells us what people are looking for. Enter content production and the art of building landing pages that convert, which gives us a means of engaging visitors with copy, audio, and video. Enter email marketing and social media, two methods for keeping up with our prospects between the point of conversion and the point of sale.

Remember when I said that neuroscience is still a bit of a guessing game? Remember when I claimed that an entrepreneurial background is more valuable than a PhD in statistics? This is why: All of the statistical analysis in the world won’t help us if we can’t apply that knowledge to creating better content and improving our offer.

Marketing analytics is a lot to learn and a lot to put your trust in but there is only so much to test and analyze from a mathematical perspective. There is way more to test from a creative standpoint but you do have to have the technical side in place first. And you have to know what measurements apply. There are only so many functions to test, track, and improve: click-through, goal conversions, and sales. More general metrics like time-on-site and average page views are helpful too but in the interest of growing sales, we want to focus on the former over the latter. It is a process of continuous improvement and we must keep in mind that what works best today is often different than what will work best tomorrow.

Marketing analytics seems to always keep going and going and evolving and changing but if you are not focusing a portion of your marketing budget on analyzing the return on your marketing investments (expenditures and performance), consider yourself antiquated in most markets. Without the proper digital marketing tools you are at a serious disadvantage but the tools alone don’t buy results. The tools merely give us a sporting chance, beyond that we must develop all of the skills mentioned above.

The learning curve here is steep at best. An individual employee cannot learn this all in a year or two, but with the right help any company can set up a properly-tracked marketing funnel in a matter of months. So it’s not easy getting started but it is worth the effort. It takes patience to build a marketing funnel with a reliable monthly ROI. Also keep in mind that marketing analytics is a custom build scenario but it needn’t be overly complicated or high tech to be effective. Sure, we can track how people scan a page and we can count cans on shelves and we can even take live video of our visitors browsing our sites and generate heat maps but that will only get us so far down the path to dollars and cents by itself. Success is still dependent on positioning, content, the offer, and the follow-up. Testing helps us determine what works and what doesn’t along the way.

Related Article: No Marketing Technology on the Planet Will Make Your Business Profitable (By Itself)