Of all the challenges companies face when they decide to build Multi-Channel Marketing Analytics capabilities, the first challenge boils down to one question: How do we start?
Here are some considerations for keeping your multi-channel marketing analytics efforts on track from the start:
1. Begin with business goals. Analytics initiatives are more likely to produce useful insights when you begin with a specific business goal or use a specific case (such as finding ways to optimize multi-channel marketing programs by predicting a customer response rates). While traditional reports and dashboards can show marketers how many customers select a particular channel, analytics can explain why customers prefer one channel over another, what type of devices are most critical for to drive the channel experience and even what is a most effective channel mix, allowing marketers to formulate more targeted media strategies.
2. Establish Analytics CoE. Many multi-channel analytics efforts are connected to IT organizations. Leading companies have begun creating Marketing Analytics CoE, staffed with multi-channel data analysts, data scientists and IT partners, to oversee disparate projects. Companies typically design the Analytics CoE to drive widespread use of analytics across the enterprise, standardize tools and platforms, and confirm multi-channel data accuracy and consistency.
3. Apply agile methodologies to multi-channel analytics projects. Many organizations set up analytics capabilities for marketers following a traditional “waterfall” development approach. They take requirements from marketers, and months later they return with a “dashboard”. The waterfall method doesn’t work for multi-channel analytics projects. Because the questions multi-channel analytics attempts to answer are forward-looking and related to new technology and channels; therefore it’s hard to know exactly what data structures and reports to build to answer those questions. Also, many marketers experience difficulty defining their data needs. The waterfall method limits marketer’s ability to change what they want during the project.
4. Start small and scale. Do not hoard multi-channel data. Don’t make the mistake of trying to gather all imaginable channel data to feed a multi-channel analytics project prior to its start. The main danger in hoarding data: by the time you have gathered the data you think you’ll need, you’ve run out of budget and time to do anything with it. An effective approach would be for multi-channel data strategists to source data according to a specific marketing strategy and business questions. Try to
5. Avoid perfectionism. Focus on rapid data integration via agile methodologies. Multi-Channel Analytics is about finding faults quickly and intelligently so you can recalculate your models faster. In many cases, you have to run through several iterations of the model algorithm before it’s correct. If you use a waterfall approach, your budget will run out before you deliver an actionable insight.
6. Create a sandbox. Traditional BI reporting tools aren’t fully adequate for the world of multi-channel analytics. These tools don’t permit the open-ended data manipulation and discovery required to answer forward-looking marketing strategy questions. You should look for more flexible applications like Tableau or QlikView to complement your traditional data collection tools.
7. Test, test and test. Multi-channel analytics isn’t about reporting or building systems. It’s about assembling small teams that can conduct quick studies of specific business problems. To answer business questions using analytics, Analytics CoE needs to come up with hypotheses, test them against different data sets, and revise models as required. The test-and-learn approach is much different from the waterfall method.
8. Work on a communication plan. Multi-channel data modelers and analysts can do more effective work when they maintain an ongoing dialog with marketers and key decision makers for whom their work is intended. Frequent, two-way communication helps reduce the risk of unfortunate downstream surprises, expensive implementation delays, and speeds up the end-user adoption at the end of the project.
9. Balance “real world” and “data science”. Analytics is a type of science, and therefore textbook knowledge is crucial. But it shouldn’t trump the ultimate business objectives. Sometimes, skilled analysts can be overly confident in their abilities and in the accuracy of their judgments and models.
10. Analytics governance and education drives the end-user adoption. It’s not uncommon for some marketers to be skeptical of the value of multi-channel analytics. Their skepticism frequently comes from misconceptions about analytics and what it can and can’t do. Because multi—channel analytics is sometimes mistaken for off-the-shelf software which can “predict the future,” some marketers believe that analytics solutions will provide them with a kind of absolute truth. This can lead to unrealistic expectations—and disappointment—when analytical models fail to provide absolute truth. The goal of any analytics project is not “absolute truth”, but to convert raw data into insights, inferences or predictive models that can lead to better multi-channel marketing decisions. While analytics governance continues to evolve as a discipline, I see it as a process by which marketing and analytics organizations define and manage different types of multi-channel data and insights. Synonymous with “quality”, analytics governance strives to ensure marketers have reliable and consistent data sets to assess multi-channel performance and make decisions. It is critical to remember than effective analytics governance is not a one-time exercise, but a fully developed effort and repeatable process.
Overcoming these technical, behavioral and organizational barriers can be beneficial in improving the returns your marketers can get from multi-channel analytics initiatives.