Six Steps To More Actionable Marketing Analytics

Numbers will never tell the whole truth. But when interpreted holistically, they can provide the basis for analysis that leads to more informed decisions and ultimately to better results.

Numbers don’t lie, but they don’t tell the whole truth. This reality is all too familiar to marketers, who pull seemingly endless reports from multiple sources, only to have the data say different—and sometimes conflicting—things.

Let’s take something as “simple” as daily revenues. With a first glance at two similar reports, it’s clear that the reports don’t look the same. Worse yet, they don’t match. In this case, the total daily revenues in two different reports might show two different numbers. This can happen for a number of reasons:

1. One report is for new customers only, and the other is for all customers. (I literally just came across this yesterday, and it was only because of a random question that I realized it.)
2. One report is gross sales, and the other includes refunds that were processed yesterday.
3. One is from an outside vendor, and one is from an internal resource, but their definitions for what they are pulling are different. (Sometimes this is intentional to make their numbers look as good as possible.)
4. One report is based on cash billings, and the other is based on GAAP rules.
5. Yesterday’s numbers change based on when you pull them; those familiar with Google Analytics and Adwords know what I’m talking about here.

The list could go on for pages. Regardless of how data-centric your role or organization is, it’s frustrating when two reports don’t match. And typically, if two reports don’t match, there is a good chance numerous others don’t as well.

You want to be a more quantitative marketer. But these at-odds reports seem like more harm than good. Don’t despair. It’s possible to have an analytics program that provides real value, in a reasonable amount of time, with minimal back-and-forth. Here are six ways to get started:

1. Define your reports clearly: Start by gaining clarity on what a report is supposed to mean. This is easier said than done. Ask yourself: If the data doesn’t clearly roll up to a defined KPI, who cares? Once you’ve defined a report’s purpose, label it prominently as such, whether in the header, footer, or somewhere else. You don’t need 37 footnotes in each report, but you shouldn’t assume that “people just know what this report is because they’ve been looking at it for a while” either. Find a healthy balance between precise definition and annotation overkill.

2. Validate across departments: Once your reports have been defined, share them across departments to ensure that they meet everyone’s definition of truth. This requires collaboration between marketing, IT, and analytics. In some companies that’s three different people; in others, it might be one, but hopefully it’s not zero! When all involved parties have signed off, validate with your stakeholders. Make sure that the people running the reports are on the same page and the people reading the reports know and value what they’re looking at.

3. Audit your reporting tools: Businesses often have too many reporting systems in place. Each department likes to pull “their” numbers from “their” system, and even then, different people in the same department sometimes pull from a different system because that’s what they are used to. The reports are then kept in Excel, Google Docs, on a network drive, or on someone’s laptop. Determine which reports you actually need and which tools provide them. Standardize on solutions across the organization, and streamline your stack to only the systems that you really require.

1. Designate an owner: Now that everyone is on board, decide who owns your metrics. This is often the trickiest step, fraught with organizational, technological, and political challenges. But while it’s not easy, it’s critical: Successful analytics programs are almost always centralized.

I work with several companies that have multiple reporting tools in place or at varying levels of implementation. None of them has a true owner of reporting, analytics, and data integrity. Several people are involved in these projects, so it’s not for lack of resources, but it’s also about the right resources being deployed. In my opinion, the owner should be on the marketing team or analytics team—but if the latter, he or she has to be very closely aligned with marketing.

5. Eliminate bias: This often goes hand in hand with the steps listed above, but deserves its own call-out. Sometimes disparate versions of the truth are the result of people’s (natural) desire to make their numbers look as good as possible. For example, when it comes to attribution of orders back to media, someone may use first click versus last click attribution to report their numbers. Others may include view-throughs as well. (Don’t get me started on that one). If you layer on how to map offline media to online orders, things get even messier. These issues can be made worse when the TV folks report to different people than the online media folks or when the SEM person reports to a different part of the organization than the person running Facebook ads, even though both are focused on digital media.

Defining and validating reports and designating an owner go a long way toward mitigating these issues. But it’s cultural as well. While it’s important for employees to be held accountable, make sure your data-driven ways don’t drive people to put their interests above the integrity of your data.

6. Consider outsourcing: Marketing data and analytics are only getting more complex. It’s okay–and frequently preferable–to keep your core competencies in-house and to outsource the rest. Analytics and attribution experts abound. You have options. You can hire an end-to-end third party service provider to pull and report on your numbers. Or bring in a strategic technology partner to help develop analytics and attribution models most applicable to your business. Many platforms streamline centralized, consistent reporting in a media-agnostic (and employee-agnostic) way, removing roadblocks at each point of the reporting process.

Caveat: Even if you outsource, someone internally still needs to own and define the metrics. And ultimately, you and your team will (and should) understand your business better than any third party; how you interpret the reporting and the actionable steps you take as a result of the analysis must remain an internal accountability. The ownership decision, however, becomes much less politically charged when the primary responsibility shifts to an external partner.

At the end of the day, props to those of you who are making strides to become a more data-driven marketer. But if you have poor data integrity, you’re only a bit better off than those who don’t look at data at all. If you’re lucky. Sometimes bad data is worse than no data. Numbers will never tell the whole truth. But when interpreted holistically, they can provide the basis for analysis that leads to more informed decisions and ultimately to better results.

See what the Twitterverse is saying about big data: