Top-down Vs. bottom-up: Integrate the ‘Why’ into analytics

You are here:

Using a bottom-up model to guide your analysis, you can gather amazing insights from the data you already possess

The key to data analytics lies in asking the right questions, but success can often hinge on the person who is asking those questions.

Traditionally, organizations have analyzed their businesses from the perspective of the boardroom, approaching problems from the top down. But in trying to address how to best use their data to solve problems without first understanding the data itself, errors of human bias and oversimplification of the issue may occur. Consequently, company leaders may invest time, money, and resources into technology and products that don’t deliver on expectations. Just because a software package comes with a flashy dashboard and a colorful knack for highlighting correlated relationships doesn’t mean it’s all that great at proving the causal ones.

But there’s no turning back now, and there are a number of valuable ways marketers are using data to advance the growth agendas for their firms, including predicting customer behavior, quantifying the return on the customer experience, and redesigning loyalty programs — just to name a few.

But who decides which problems to tackle and why? As the competition for organic growth intensifies, the need for all businesses to become more data-driven increases. An executive leadership mad dash ensues that all too often doesn’t include or respect the voice of the data scientist — and that’s a big mistake. Given the high cost of getting data analysis wrong, it’s more important than ever to take the best approach to getting your company’s marketing solutions right.

How Are CMOs Getting It Wrong?

Today’s marketers need to be creative and analytical. CEOs now expect their CMOs to grapple with all the technological avenues out there to reap the rewards of data analytics.

Consider a respected marketer who doesn’t possess the skills or knowledge required to achieve optimal results using data yet still believes he’s the best person for the job. When marketers’ egos won’t allow them to take a backseat and hand the reins over to a qualified expert, a tendency to rely on oversimplified technology can lead to a colossal data letdown. Innovative methodologies and technologies have made finding elegant solutions through data so simple. But a lack of analytical insight on the part of those making the big decisions could be the No. 1 reason many companies aren’t effectively employing data analytics to improve their vast marketing potential.

And a dearth of deep analytical talent is only going to make things more challenging. One recent study found that the U.S. stands to see a 50 to 60 percent divide between the number of good data scientists available and the number that will actually be needed by 2018.

If you’re lucky enough to have the ability and the wherewithal to hire a good data scientist — or, even better, a team of data scientists with a cross section of skills — you need to know how to use him or her. Even successful businesses may need to alter their cultures by embracing and promoting the ideals and beliefs of their data analysts and holding them in as much esteem as they do their qualitative counterparts.

Changing your organization’s outlook is a big decision that requires a lot of internal education to get all stakeholders on board.

The Solution Lies in People, Not Software

It’s actually pretty simple: If you want to get the most out of your data analytics, find people who understand not only how to analyze the data, but also how to collaborate closely with marketers to uncover which data is most valuable and why. Consider knowing what the five most important data relationships in the company are in order to achieve marketing impact. Then, make your decisions about key initiatives and investments to yield the returns from that knowledge. That’s what’s possible with a bottom-up approach.

The tricky part is establishing the strength of these data relationships, which is easier for data scientists, who are accustomed to getting their hands dirty in the data and growing intimately familiar with it. These are the key players who are able to understand the metadata behind it all, the coding within it, what it takes to extract it, and what the circumstances were when it was collected. A good data scientist can also discern the questions the data is posing, as well as where the data points are leading, in order to come to educated conclusions and make more accurate predictions.

This puts data analysts in the perfect position to explicitly state what needs to be done. Marketers can then shape those ideas into workable projects by sourcing information and taking the actions needed to reach the predicted outcomes. The movement begins at the bottom. As the causal relationships and concepts your data scientists illuminate ascend through the organization, all stakeholders become inspired to jump on board. In this way, even the loftiest of goals can be achieved.

By identifying the lucrative resources that already dwell within your data, you can develop successful strategies within a short time frame. And working with talented data scientists enables you to make solid cost calculations, decide how to best integrate the data, and even make plans for collecting further data on an ongoing basis. All of these are crucial factors in being able to form a realistic plan.

Building a system of strategic analytics that can decipher what the data is actually telling you will help solve your marketing problems and enlighten the boardroom in the process.

Don’t Forget to Make It Relevant to the Sponsor

If you’re a CMO who needs to win over the support of the organization for such an initiative, remember to make it relevant for the department or sponsor. In the abstract, it’s really wonderful to think about finding the 5 or 10 or 20 most impactful data relationships in the business, but most executives don’t have the latitude of responsibility to follow the yellow brick road to wherever it objectively leads. Ultimately, they’re paid to get sales, finance, or marketing working at their very best. The CEO, chief strategy officer, and chief operating officer may prove exceptions to this rule. But if you’re not undertaking this analysis in collaboration with one of those executive sponsors, you’ll likely need to confine the investigation and the results to just one department.

But don’t forget to build on your success. As you make key data relationships that exist in your area of responsibility actionable, advertise your success, and you may soon win the right to travel across the entire organization as the one executive making the biggest impact on the firm’s growth agenda.

Ultimately, smart marketers who partner with talented data experts will free up their own time and be able to direct their creative talents toward strategizing for the future rather than generating reports about what has already taken place. The real value lies within. Using a bottom-up model to guide your analysis, you can draw amazing insights from the data you already possess.

John Kelly leads the predictive analytics practice at Berkeley Research Group, which works with marketing, sales, and operations leadership across a range of industries to leverage the power of econometrics and data science. This work results in evidence-driven management, delivering dramatic growth and performance improvement. Some of the specific ways it helps clients include dynamic pricing optimization, loyalty program design, and predicting consumer behavior.

From our sponsors: Top-down Vs. bottom-up: Integrate the ‘Why’ into analytics