Data-Driven Marketing — Do You Really Know What You’re Getting for Your Digital Ad Spend?

Digital ad spend is now a $300 billion a year business and growing, with Google and Facebook grabbing over half of that. That’s a lot of money. And as spend has increased, so has the demand for more granular and sophisticated targeting. In addition to their large user bases, one of the reasons Google and Facebook are way ahead of the pack is their ability to leverage machine learning to continuously improve the targeting ability they offer to advertisers.

You’d think, then, with so many dollars at stake and so much data available, that anyone with a decent-sized marketing budget would be obsessed with maximizing the return from their digital marketing. But as I talk to different companies out there, here is what I have observed:

· Almost everyone knows how much they’re paying for clicks and converts

· Some know how their “converts” translate to true sales — but not as many as you’d think!

· Almost no one knows how much true return they’re generating from each sale

Why is that? Because very few companies understand the life-time value (LTV) – the net profit of the entire future relationship — of their customers at a granular level.

The 80/20 Rule

Like many things in life, not all of your customers are created equal. If you’re like most businesses, the Pareto Principle will roughly apply — 80% of your profits will come from 20% of your customers. For those 20%, it stands to reason that you would be willing to pay more for a sales convert — and the inverse for the other 80%.

Here’s a quick example. Let’s assume that Company A spends $10 million on digital ads and generates 100,000 sales, which they earn $20 million in LTV from. That’s a 100% return-on-investment (ROI).

But those sales probably break down something like this:

The only segment they are actually generating positive ROI on is Segment 1! If Company A can adjust their bidding strategy to allocate more marketing budget to prospects in segment 1, they will increase sales to their most profitable customers and generate higher ROI. Of course, that’s only possible if they can:

a) predict what characteristics determine who will be a Segment 1 customer, and;

b) execute their ad bidding across those same dimensions

And that’s where things start to get a little tricky.

Why Aren’t We All Doing This?

The first reason is data.To build effective LTV models requires historical campaign and customer data that is clean and accessible. Companies that have been around a long time will undoubtedly have data stored in a vast array of locations — some of it on mainframes, some of it in databases — with little centralized knowledge on how to bring it all together. But data is like oxygen when it comes to performance marketing. Investing in data quality is a prerequisite to building LTV models to drive your marketing strategy.

“But what about all that data I get back from my ad provider?” Unfortunately, this is the one area where Google and Facebook can’t help you. They have lots of tools for optimizing your conversions — Google even lets you import your own conversion data — but they can’t access your customer performance data post-conversion. Only you will know that — how many future purchases that conversion has generated, what is their average basket size, how has their payment history been with you.

The second is organizational design. Large companies are rife with siloes that are organized by function and do not effectively coordinate on integrated strategies. The analytics team holds the data, but they don’t talk to marketing. The finance team owns the models, but they care about vertical returns and not LTV. The marketing team owns the budget, but they’re worried they will lose it next year if they don’t spend it all. And so on, and so on. To effectively build and deploy LTV models requires alignment among all these areas if it’s going to be done right and move the needle on your marketing return.

The third, and often overlooked, is culture. I mentioned before that data is like oxygen. Is this true in your company? A commitment to a data-driven marketing strategy starts with a truth-seeking culture. We don’t seek out data to prove our ideas; we search through data for the ideas themselves.

I learned this early in my career when I was an analyst at a large bank. My manager asked me to conduct an analysis to support a recommendation he was bringing to the President. I wanted to impress him, so I conducted a very thorough analysis and came back with the right answer, clearly evidenced by the data. Just one little thing, though — the data showed that the right answer was the opposite of his recommendation. What do you think happened next? You got it — he immediately dismissed my analysis as “wrong” and proceeded with presenting his original idea to the President. I quit that job shortly after and vowed to never again work for a company that doesn’t have a truth-seeking culture.

I’ve seen a similar flavour to this when it comes to marketing decisions. Even though the models say to switch things up, the CMO will want to continue to “go with their gut” based on what they believe has worked for them in the past. Building models isn’t always the hardest part; actually doing what the models tell you to do often is!

Evolving Your Marketing Strategy

So how is your company doing when it comes to getting the most ROI from your digital ad spend? Here is a quick checklist that can serve as a barometer.

For your organization:

· Do you have statistical models that predict key outcomes such as future spend?

· Do these statistical models feed into LTV models to measure the net profit by segment? How granular are they — Product-level? Model score segment-level? Customer-level?

· Do you refresh these models frequently? Quarterly? Monthly? Weekly? Real-time?

· Can you translate these model predictions into actionable bidding units, such as keywords for Google or demographics for Facebook? Do you?

· Do you conduct A/B or foundational tests to enrich your dataset? Do you systematically review test results and adjust your strategy accordingly?

· Do you share campaign results with a broad audience? With the CMO? With the Product Leads?

· Do you consistently measure total ROI and use that to determine your marketing budget?

Being able to answer ‘yes’ to all of these questions requires a tremendous investment and commitment to a data-driven strategy — but one that is well worth it and can put you well ahead of your competition!

About the author

After a long career as an executive in financial services, I recently started my own company, Payson Solutions, to help companies transform their business. I have a passion for building high-performing teams and leveraging advanced analytics to build amazing customer experiences. If you would like to connect, drop me a line at brent@payson.ca.