In March 1941, British cryptographer Mavis Batey broke the Italian version of the Enigma code and vectored a British battle group towards an Italian force near Cape Matapan, at the southern tip of Greece in the Mediterranean Sea. At 8:10 PM on March 28, the British cruiser Orion used a then-cutting-edge innovation called “RADAR” (Radio Detection and Ranging) to locate the Italian ships at night in heavy fog. A battle commenced, with the British achieving surprise and winning a significant victory. Both forces engaged in the Battle of Cape Matapan were formidable. Superior data gathering and interpretation followed by decisive action made the difference in a struggle that could have gone either way.
The stakes we compete for in digital marketing are incomparably lower, and our technologies far more advanced, than those of our foremothers and -fathers, but some of the same decision principles apply. The Brits invested heavily in gathering, collating, analyzing, and interpreting data, at multiple levels of their command structure, at a time when they were counting every penny and pound. They knew the value of sound, well-informed decision-making and put many of their resources into insight generation.
Modern marketers compete against savvy competitors to attract and convert fickle consumers in a landscape of always-shifting technologies and tastes. Most steps are wrong steps, in the sense that it is very hard to perceive the exact point at which the amount spent on any channel, tactic, placement, campaign, ad group, or keyword exceeds the marginal profit that spend drives. It is far more difficult for those marketing channels (such as network television) where there is no digital trail of consumer footprints leading from the marketing cause to the conversion effect.
President John F. Kennedy said that “Success has a hundred fathers.” Online conversions have at least that many, with half a dozen digital tools giving contradictory opinions as to which channels and tactics should get what transaction credit. DoubleClick, AdWords, and Bing Ads look all the way back to the very first impression that each of them served to award the lion’s share of conversion credit to display and paid search, respectively. Site-monitoring tools like Google Analytics and Adobe Analytics typically start tracking (and crediting) from the inbound click. They have a better view of traffic driven by all channels and the total online conversion pie but do not typically give “view-through” channels credit for paid channels’ upstream achievements. And their first-click, last-click, and various other non-algorithmic credit-spreading schemes are but dartboard guesses at the true causes of the conversion goodness that is happening on your site.
For a “few” dollars more, attribution specialists are willing to take your data (assuming you’ve got it available in a centralized data warehouse, more on this later) or, in some cases, generate their own data by tagging your site to generate a more accurate, algorithmic estimate of the fractional transaction credit each channel and tactic deserves. But the high costs of these options can outweigh their benefits. One can afford a lot of bad marketing spend for the price of an attribution model, and if the Orion had offloaded all of its ammunition to make room for its radar before putting out to sea, one can imagine the battle having had a different outcome. If you aren’t spending a certain amount on marketing per year across multiple digital channels (among which I include the P, E, and S elements of the PESO taxonomy), the attribution vendors themselves will typically wave you off. Cheaper attribution options exist, such as Google 360, but there is an inherent conflict of interest in getting one’s attribution from the very people who own the digital channels.
It used to cost hundreds of thousands of dollars to house the millions of data rows necessary for attribution modeling and other digital decision-making…dollars the clients of enterprise data warehouse vendors like Teradata and Oracle gladly paid because they understood the improvements in decision clarity these data repositories enabled. Then came Hadoop, which commoditized data storage for those with programmers talented enough to set up Hadoop clusters and patient enough to wait for their snail-paced queries to complete. Now, cloud-based providers like Amazon Web Services offer the best of both worlds, enabling the rest of us to set up reliable, scalable enterprise data warehouses from which regular, SQL-speaking analysts can quickly and easily retrieve data that can be flushed to decision-makers via customized business intelligence dashboards, or shaped into laser-guided decision documents, all from an investment of couch-cushion change (depending on one’s couch).
Should you? Andrew McAfee and Erik Brynjolfsson discovered in 2012 that “the companies in the top third of their industry in the use of data-driven decision making were, on average, 5% more productive and 6% more profitable than their competitors.” Setting aside the organizational obstacles mentioned by Chuck, these gains follow substantial investment in talented personnel. So, ask yourself what it would be worth to know precisely:
- How much profit you make from the incremental sale of the very next widget
- How valuable your customers will prove to be over the lifetime of their relationship with you
- Whether site visitors convert better with the green buttons or the blue ones
- How much to spend in what ways on marketing
If your company is in a stable industry, insulated from competitors by patents or trademarks or differentiated skills, with strong market power over the firms up- and down-channel then the answer to this question may be “not much,” and that’s okay. Please accept our envy.
If, like the Cape Matapan combatants—you scan the horizon with trepidation, then you’re like the rest of us. But hopefully, now you’re a bit more informed.