top of page

Data-Driven is Not Enough

Digital loves buzzy words and expressions, and one of the most popular is being “data-driven.” Of course, the idea of leveraging data to drive better decisions is simply too smart not to embrace. But I am not convinced that being “data-driven” is enough.

Fifteen years ago, I was working in business development for an agency. At an industry event, I spoke with a brand leader who was new to digital and lamenting how his digital team and agency were missing the mark on being data-driven. He summed up the issue entirely in four words:

They tell me what but not why.”

The business leader knew that having access to loads of data – even important data – didn’t, by itself, help that company grow. This marketer – and all savvy leaders – wanted to understand the forces behind campaign data.

Here are three examples that show “whats” that need “whys” to be actionable.

  • A particular partner class drives more sales than others and does so at a tremendous scale.

  • Over the last two months, sales through the top partner declined 30%.

  • The team loved the idea of partnering with this high-quality content partner, but they aren’t driving sales at a volume they had hoped.

Each is a set of data-derived facts that one could accept at face value. But the real actionability comes from the whys. The following real-world stories show how data-driven analysis and decision-making dramatically improve business results.

A particular partner class drives more sales than others and does so at a tremendous scale.

When a partner class delivers lots more sales than others, many people will think, “gotta do more of that.” But one footwear company I am familiar with dug deeper to understand whether that partner type delivered the best business value. They looked at the percentage of sales to new customers, purchase frequency, and average order value to learn more.

They found that virtually none of the “hot” channel customers were new. When paired with data on average purchase size and purchase frequency analysis, the data showed that the sales were of much lower value than those through other partner classes. Other seemingly more expensive channels drove up to 90% net new customers.

This company adjusted its commission programs so that the bounty for partners reflected the incremental business value of each sale. They kept their relationships strong with their top partners in the “hot” channel by saying, hey, you’ll get less per sale, but we’ll commit to giving you the same revenue for the year if you work with us. Most of those top partners stuck with the brand as it changed commission programs to reflect customer quality. Once they implemented the changes, they could prove that they were driving far more sales and margin growth than ever before and were driving more incremental revenue and profit than other channels. This changed channel perceptions among senior company leadership, who agreed to invest significantly more in affiliate.

Over the last two months, sales through the brand’s top partner have declined 30%.

Reporting tools are great at visualizing critical data, but they usually cannot surface why things happen. That was the case in this example from the accessories industry. The team investigated. There was no immediate drop in performance, so it wasn’t just a broken tag or tracking failure. The decline was steep but gradual. Program commissions and terms had stayed the same.

By digging deeper, the brand was able to see that while the one partner’s performance had dropped, similar publishers were doing better than ever. Further, they analyzed competitor programs on the affected site and found no major changes in programs or commissions during the period.

They met with the declining partner to uncover the whys. They shared sales and click data for the entire partner class, and the partner reciprocated with their category revenue. The partner’s team realized that category sales were dropping across all clients through this data sharing and collaboration. By analyzing the journey for the customers that clicked, they found that they had inadvertently reduced the “real estate” devoted to the category. By walking back the changes, both the publisher and brand won. It was all enabled by finding insights for and from the data.

The team loved the idea of partnering with this high-quality content partner, but they aren’t driving sales at a volume they had hoped.

A tier-one content partner was gleefully added to a sporting goods program I am familiar with. The team had high hopes that the site’s massive reach would create a new top-5 or -10 partnership while making internal leadership more excited about the channel.

After four months, the partner drove decent traffic and sales but did not reach the top 10 – it squeaked into the #19 slot. Why the disparity between expectations and actuals? The team went deep into their data in collaboration with the content partner to understand more. First, while their content partner’s site traffic was enormous, traffic to performance sections was modest. Second, their program drove many clicks but a much lower conversion rate. Many would-be sales were ultimately credited to other partners because of toolbar interception and last-minute CTRL-O coupon searches just before checkout.

How did the brand change? First, they right-sized expectations. Then they discussed new ways to drive additional traffic to the promotional content. Finally, they adjusted their payment model to include a modest bounty for first clicks to improve partner revenue. Results improved.

So What?

There’s nothing magic about these examples. The “sorcery” is in the interest and willingness to understand the forces behind the data. The best brand leaders are insight-driven. Machine learning and algorithms are better at reacting to data than a person can be in most cases. Great program managers excel in digging deeper to uncover the meaning behind the numbers.

Access to data is nothing without insight. Data-driven insight is what we need. The other lesson from these examples is that insight-driven doesn’t necessarily require new tools or platforms. Great insights tools and platforms are great, of course, and can help you uncover insights far more advanced than the examples above. You don’t need a Ferrari to get from point A to point B. What matters is being willing to go from the data to the insights the numbers help reveal.

On the Record

Be Informed on the Regular.

Receive news, premium content, and event alerts right in your inbox.

bottom of page