Getting Big Data Right

Data, data, data. That’s pretty much all anyone hears nowadays. Yes, data is critical to modern marketing — and will only grow in importance as all media becomes digital. But not all data is created equal, and figuring out a winning data strategy isn’t a slam dunk. Here are five guidelines for success

Not All Data Is Useful. As marketers, we must identify what data is important to run our businesses and manage campaigns effectively. We should be setting priorities to what data needs to be immediately included in our analytics and campaign management versus data that is nice to have. But to do this right, we must develop a centralized data selection and sourcing process enterprise-wide for the marketing team. Evaluate data based upon quality, collection methodology, granularity, availability, and frequency of refresh. As marketers, do we need to collect cookie-level data or will summary-level data suffice for our needs?
Peel the Onion. One approach we use for data management success is what we call “peeling the onion.” That is, we start with what we want reported and then determine the data that is available to support that analysis. Don’t let the data guide the reporting lest you’ll never see the forest through the data trees. For example, a large brand might not need to know the sentiment of each and every blog post, but it may need to know the collective sentiment of posts by each market daily. Make decisions on what data should be collected based upon reporting needs not data availability.
Keep the Insights Not the Cookies. As future legislative action takes place, the risk to storing historical insight based upon cookies — even anonymous cookies — becomes ever riskier, both in the U.S. and Europe. At minimum, marketers need to develop data marts that can maintain the insight of all historical insight in a language that is useful for long-term analytic needs. Also, marketers must determine their process and strategy for anonymizing all their marketing data to still be useful across the entire marketing organization.
Centralize Your Data Management. Gathering data in a timely and efficient manner is one of the most challenging aspects of managing big data. Marketers need to centralize their data management processes and tools to develop accountability for data updates, maintain a handle on data deficiencies and alert the data users for data holes. Rating systems have hiccups, ad servers go down, data APIs get changed without a centralized data management strategy and systems; marketers will spend way too much time managing data collection and transformation. Shouldn’t analysts spend 80% of their time optimizing marketing rather than hunting down data set A from vendor B?
Develop a Transportable Data Model for Analytics. Since marketers have many stakeholders involved in the analysis of marketing data across various teams (e.g., research and analytic firms, internal teams, and agencies), marketers must develop a data mart that is transportable across multiple analytic and reporting systems. Without a transportation model, a marketer will have multiple versions of the same data floating across their stakeholders. This will negatively impact our ability to make timely and accurate business decisions. Never mind the operational challenge of having to refresh all these data sets daily.
Following these won’t guarantee success, but it will give you a much better shot.
Adam Gelles, CEO, The AMM Group.
https://digiday.com/?p=4101

More in Media

Why podcast companies are investing in AI-generated podcast translations despite questionable quality

While podcast networks like iHeartMedia, Spotify and PodcastOne have publicly announced plans to debut AI-generated audio translations, few have gone live yet.

WTF are OpenAI’s custom GPTs?

OpenAI’s custom GPTs and GPT Store could proffer a potential new platform for publishers to drive referral traffic.

AI Briefing: Elon Musk and Sam Altman battle over OpenAI’s mission

Legal pressure on AI companies illustrates the myriad challenges for companies that want to use or build generative AI tools.