Guest Post: How Broken Is Your Retail Marketing Data?
By Madan Baradwaj, co-founder and CTO at Measured
The era of “big data” was supposed to change the game for marketers.
The experts said big data would help marketers better understand their customers, how to reach them, and ultimately get better results for their companies. In some ways that has happened; but at the same time, more data has simply left marketers drowning in information without being able to use it effectively. The old saying about not being able to see the forest for the trees comes to mind here.
But data quantity is only part of the problem; the real challenge is with data quality. In its 2018 marketing data report, Dun and Bradstreet found 89% of marketers believe data quality drives the right sales and marketing campaigns, but only 11% were extremely confident in their companies data quality.
Besides gathering customer data for building campaigns, marketers also rely on data to measure the success of those campaigns. For a decade, data quality has been the bane of cross-channel marketing attribution analytics, and today it is getting worse.
Multi-touch attribution promised to use data to map what advertising touchpoints led to customer conversions. However, marketers found themselves spending tons of time cleaning data only to find out that couldn’t get any insight from it in the end. The problem is pretty simple: when you collect data by pixels, you can a great sample of demographic data, but there is no source of truth for how the customer moved through the buying process or which ads they saw along the way.
Marketers therefore ended up with broad assumptions but nothing to inform a decision. And by the time, the data was scrubbed, the campaign was over.
If this were only a data quantity problem, the engineers could solve it with bigger pipes and more capacity, but data quality is much harder to solve. There is more diversity in data coming from pixels, CRM programs, social media, and on and on. These different sources produce data in different formats so analyzing it together is difficult. Now with new challenges for data access because of the EU general data protection regulation (GDPR), limitations on third-party tracking and Facebook’s unwillingness to share marketing data, the problem is worse.
So how can marketers test their data to know how broken it is? There are a couple straightforward tests. The first simple one is to open your Google Ad Manager and Google Analytics. Count the conversions and see if they match.
Want to know about Facebook? Check the conversions that Facebook counts in its analytics tool and compare them your Facebook conversions in Google Analytics. You can also check the total Facebook clicks vs. what Google Analytics counts as your visits sourced from Facebook.
I’ll bet that you’ll see big data discrepancies between these analytics tools, and it will prove you need a better way. The best method for measuring the effectiveness of your marketing campaigns is to look at them more narrowly with good data science-based experimentation.
Start small and A/B test incremental data from your campaigns against metrics that matter, like revenue, and see which ones had an impact. You’ll get answers right away, and in time to adjust the campaign before it’s over so you aren’t throwing good money after bad.