3 ways to make your data reveal its secrets
In an age when big data is everywhere, and AI becomes part of the marketing vocabulary, many marketers are reluctant to admit they still struggle with making sense of their data. That problem is, however, commonplace.
I recently interviewed about 15 CMOs across a range of B2C industries, including heavily regulated ones such as healthcare and financial services. I noticed a marked gap between those who mastered their customer data flow, were able to derive real-time insights and deliver impactful personalized marketing, and those who were still trying to stitch together various data sources and manually run excel reports: “We still don’t have simple, yet effective tools to handle data” lamented one.
A just-released survey from Deloitte and the American Marketing Association confirms the challenge: an alarming 68% of marketers don’t use analytics data to help make daily business decisions! Complaining from overly complex and poorly integrated tools, these marketers plan on quadrupling their analytics data budget over the next 3 years to remedy those difficulties.
As you hone your marketing competitive edge, here are 3 ways to make your data ‘sing’:
Take a fresh look at your data
“The amount of information you can collect on an individual is pretty remarkable… once we have your mobile ID, we can use that to collect information from multiple sources” admitted a savvy CMO. To derive new insights from this data trove, avoid confirmation bias: “I think it’s critical to be somewhat naive when you look at the data that you’re seeing. You can always find data to support some supposition”, advises John Koetsier, Mobile Economist at Tune. “I will just look at the raw data for a while. I will start to average out certain rows and columns. I will look for outliers. I will look for patterns in the data. I’ll try to graph it, four, five, six, to seven different ways” he continues.
Data marketplaces such as AppAnnie provide a wealth of comparables: “We’ve seen other companies be really smart in terms of looking at market adjacencies and seeing what feature functions were introduced by a company. Let’s say a certain mobile payment is being integrated by a quick serve restaurant or retailer, and it correlates to an improvement in either the number of active users, [or even drives revenues]. You can use the learnings from other apps that have introduced certain features, that would be really expensive for you to bet on too early” suggests Danielle Levitas, SVP Research & Professional Services. Today, many of your martech vendors integrate intelligent algorithms that draw patterns across pooled datasets: “We extract data, train our neural networks from deep learning algorithms, and that is how AI is able to read messages and suggest responses and method data for a specific case” says a chatbot vendor.
Connect gut and brain
“While there’s clear expertise, experience, and some level of people’s innate sense of direction, the fact of the matter is that data has become more critical throughout organizations,” says Levitas. But looking at data patterns is not enough to derive insight. You’ll need to answer the ‘why’: why is this happening, what do customers want, what does this mean for our business. “I often try to get survey data as well as the quantitative data”, adds Koetsier. Meshing data points with opinions, whether internal, from customer surveys, or from industry experts is necessary to give relief to your analysis. Do not forget that being the statistic, there is a human being you are marketing to. “I think the future lies in user experience, where you can understand the user intent and preferences, and advertise to them, leading to personalization and customizations beyond the marketing funnel” concludes a CMO.
Believe me, marketers that are able to make their data sing are only limited by how creatively they can use their insights. Remember that auto loan ad that fortunately guessed you were thinking of buying a new car, even before you started searching for one? There was no voodoo magic behind it, just smooth, effective data crunching.