Two women are eating lunch in a resort. “Boy, the food here is terrible,” says one. “Yes, but such big portions!” raves the other.
Okay, that’s not quite the way that old joke goes. (The punchline is supposed to be, "And such small portions!") But my version is an apt description of marketers’ access to data in recent years. Data is available in quantities that are almost unimaginable, but, too often, the quality is mediocre.
There has been a marketing industry preoccupation with scaling data over several years, but recently the pendulum has swung back toward quality. Marketers are less interested in scale and more keen on quality standards. That’s a positive move for the industry, but the pursuit of quality raises its own questions. What does “quality data” really mean? How is the data collected? Is there an objective standard for audience data? Let’s examine these points one by one:
1. What does quality data mean?
Determining data quality seems fairly straightforward, but it could mean data that’s effective, well sourced or collected in an organized manner. In the absence of a clear definition, marketers must confront several questions:
- Labeling: What are the ingredients? What’s “in” this data set?
- Provenance: Where is the data from?
- Efficacy: Does it work?
- Accuracy: Who is it?
When they refer to data quality, CMOs are usually talking about its accuracy. Marketers are asking if the information contained inside the data files is true and correct. If the cookies say males age 18-34, are you actually reaching that segment? Quality inputs return quality results, so it’s imperative that the data marketers use is accurate. And accuracy needs to be considered relative to both targeting and execution—in other words, messages should not just be aimed at consumer segments but reach them as well. It’s not one or the other.
2. How is the data collected?
There’s a strong relationship between how data is collected and its accuracy. The process of onboarding raw data from disparate sources often results in data duplication and inconsistencies.
Pressure on vendors to scale data can lead to situations where data providers are adding attributes to as many mobile IDs and cookies as possible. Such cookie stuffing usually results in wrongful attribution and inaccurate data.
The other major hazard is bot traffic. Bots simulate the actions of human customers to perpetuate digital ad fraud, costing marketers $8.2 billion a year. Some 78% of publishers experience non-human traffic on their sites. Of course, bot traffic will water down the quality of aggregated data as well.
The problems with data collection don’t end there. Data can be falsely labeled based on user interaction with different content. If a consumer reads an article about Elon Musk going to Mars, that doesn’t necessarily mean they’re in the market for a Tesla. Yet, a data-collection firm may make that leap. Finally, because people in households frequently switch devices (like when mom lends her iPad to her daughter), 100% accuracy may be an impossible goal.
3. How does this compare to the industry standard?
Unfortunately, it’s still the Wild West for marketers who purchase data. There are far too many vendors to track, and Nielsen and comScore have not yet provided standards for data quality, though both benchmark demographic data. In the absence of a single standard, marketers can ask for transparency from their vendors about how their data is sourced and how trustworthy the sources are. The most credible vendors will be upfront about their processes.
Behind those three questions is the big one: Will this work? Is the data effective?
That depends upon a few factors, including recency and relevance. If a consumer looked at a bathing suit online six months ago, that’s not very relevant by mid-winter. To mention a now-famous example, if someone buys a toilet seat, it’s likely a one-off purchase rather than a sign of a toilet seat collector.
Marketers are often dubious about data claims, and with good reason. What they’ve seen so far has been underwhelming because there has been a lot of mediocre data at scale. The challenge is to filter out the irrelevant stuff to focus on data that is truly actionable. For now, less is more. But as data quality improves, marketers will ask for bigger portions.
Evgeny Popov is APAC vice president, data solutions, at Lotame.