Business Analytics Tips: Harness Your Data to Improve Decision-making

Many organizations that think they’re data-driven are still in first gear. How do you go from simply collecting a lot of data to setting up a business analytics function that actually tells you how to tweak your model to improve profitability?

We interviewed Travis Anderson, Toptal’s director of business analytics, to get his insights and business analytics tips on setting up a central function within a company, removing reporting bias, the importance of using data, and potential pitfalls. As Toptal’s director of business analytics, Anderson leads a team that enables data-driven decision-making by connecting business strategy with data activities (i.e., data analysis, reporting, diagnostic analytics, and data science).

Business analytics supports all functional areas of the business, including sales, marketing, finance, product, operations, and HR. Anderson brings more than a decade of experience building and leading analytics and engineering teams to drive significant business growth, including at Vivint Smart Home, Symantec, Brigham Young University, and at his startup, Mapline. He holds BS and MS degrees in mechanical engineering and an MBA, all from Brigham Young University.

What Are the Uses of Business Analytics?

Business analytics allows managers to make better and more informed decisions and can increase operational efficiency by helping managers utilize resources more efficiently and ultimately optimize the bottom line, according to MicroStrategy’s 2020 Global State of Enterprise Analytics report.

In the case of Toptal, Anderson identified four tenets central to driving our business and the lifetime value of customers:

1. Acquiring customers: using data to improve the customer acquisition process
2. Expanding footprint: understanding how to drive expansion both geographically and within the existing client base
3. Retaining customers: finding points of attrition in the customer journey
4. Optimizing costs for acquisition, retention, and business operations

These four tenets are also a way for the business to measure the ROI in data and business analytics.

What Were the Primary Challenges at Toptal?

Siloed Analytics

According to Anderson, the first challenge he faced when he joined Toptal was the transformation of the internal approach to analytics. At the time, most of the internal functional teams were carrying out their analysis. Most teams had a data analyst, and they were each doing their data work, which was mostly concentrated around reporting, analysis, and trend analysis. Even though a data culture existed and line managers utilized data in their decision-making, the setup was inefficient.

Each team had a different approach, which in turn meant that the message was muddled. Since each group had an internal data function, there was no consistency in definitions and KPIs. Management discussions often focused on reconciliation, which could be a distraction. Since definitions were different, the learnings from the data were, at times, lost.

Reporting Bias

The second issue that arose from decentralized data collection and reporting was that each team had a bias in presenting its data. Each function was selecting data to portray itself in the best light. This practice created a lack of focus and a potential lack of control.

Anderson embarked on a complete overhaul of the company’s approach and business analytics framework. The priority was to create a central function: an analytics center of excellence that exists outside of business lines and serves as a control point. A central function ensures that data is collected and analyzed homogeneously and that reporting bias is eliminated.

Once the center is established, it becomes necessary to ensure that it is appropriately staffed. The first order of priority is to identify the skills gap. To build a team that can be effective and has an impact, you need people who have solid technical skills, strong problem-solving skills, but also business acumen.

How Does a Business Analytics Center Add Value?

According to Anderson, the primary added value of creating a central data and business analytics function is improving performance and reducing costs. Until a business measures performance consistently over time, it is challenging for management to improve performance significantly.

The first step is establishing the consistency of the metrics and quantifying an objective based around these agreed-upon metrics. This has the essential behavioral effect of motivating staff—as Anderson points out, how do you get people motivated if there are no goals? Furthermore, any quantitative metrics are better than none. In Anderson’s opinion, “If you only start to measure one thing, you can see a real benefit—either because you can influence it or you can see that it is not relevant.”

Anderson’s team supports all business functions and holds weekly and biweekly check-ins with each. The first part of the job is to ensure the collection of correct data. This collection serves a behavioral goal to motivate people to do their job and assign a “score” to their performance.

Choosing the Right KPIs

Once consistent and high-quality data has been collected, the biggest challenge arises: to assess what the right KPIs are for each business unit. The assessment starts from the top down. The business analytics team maps out the company strategy in data so that the selected business analytics KPIs are useful in terms of giving insights and significant on both top-down and business levels.

Some of the questions that lead to establishing the appropriate KPIs are:

• What are the key metrics?
• Are they financial?
• Are they based on operations?
• What is the framework of what the team is measuring?
• Do individual people need to be accountable for delivering specific objectives?
• How will they be rated?

It is paramount that the business analytics team thoroughly understands the business and its strategy. At Toptal, there is strong support within the company for the mission of the organization.

The data is processed and studied utilizing sound statistical modeling and forecasting. However, it is important to note that the output of the analysis is not a decision, but rather quantitative inputs that assist in making better choices. Ultimately, all business decisions are the responsibility of the business leader. There is a partnership between the stakeholders and the data and business analytics team through an iterative process. Once a decision is made, the data needs to support it. Not only, but there is a regular reassessment of the KPIs to ensure that they are always aligned with the company’s strategic priorities.

The process is not always painless. At times, there can be friction between stakeholders, as there is much feedback in the data. Not all managers are equally receptive to such feedback. Anderson sees his responsibility as providing a digestible recommendation and educating the executives on how to interpret the insights extracted from the data.

Reading the Data Wrong

Anderson touched on the potential adverse outcomes that a company can encounter when there is poor internal discipline in data collection and analysis. In a previous engagement, he had encountered a business that had a large business unit that was responsible for a substantial share of the company’s revenues. This business unit had several sales representatives who were collectively responsible for revenues of more than $200 million. However, this team measured its revenue differently from the rest of the company and reported it in a separate system.

During a management change, a new executive failed to realize that the data was not consistent and fired all team members—they had gotten the wrong insight from the data and believed that the team was not performing. The decision was taken based on faulty and inconsistent numbers in the ERP system. It ended up being a $50 million mistake. This anecdote starkly illustrates why master data management discipline is crucial, particularly for companies undergoing M&A integrations.

Common Pitfalls in Getting Started and How to Avoid Them

Anderson has encountered two typical problems in companies that begin to explore data analytics. These problems fall on two ends of the spectrum. First, companies sometimes embark upon large initiatives to collect perfect data that is ultimately not used. The second problem is when companies do not even start any analysis because of the poor quality of their data. The critical advice Anderson provides here is that even when the data is not reliable, measuring a few critical KPIs offers useful insights. Doing so will allow the company to learn how to make the inputs more reliable.

Is More Data Always Better?

While measuring the right KPIs is essential, it is worth noting that too much data (or irrelevant data) is not necessarily better. Unfocused measuring confuses decision-making and can be a distraction. It is more effective to begin by measuring a few but crucial data points consistently and correctly.

Anderson’s team’s effectiveness is measured in reference to the four tenets above: customer acquisition, footprint expansion, customer retention, and cost optimization. For each of these, the impact is measured and quantified, providing an ROI for the team’s work. If the team has done a great deal of analysis but has not inspired change, its work has been ineffective. Ultimately, the team’s success means having a measurable influence.

Anderson’s Guiding Principles for Business Analytics

Anderson’s many insights can be distilled in a few business analytics tips for the successful implementation of data analytics.

First, the mission of such a team is to change the executives’ minds through quantitative measures and to influence them every day. These will be small, incremental changes made impactful through continuous iterations and improvements.

Second, the business analytics team does not provide decisions but information that can guide executives. Business leaders are still always responsible for a company’s strategy.

Third, the impact of the business analytics function should be measurable and have an ROI.

Finally, starting with a limited set of business analytics KPIs is better than not measuring data at all. Not only, but the process creates a culture of data excellence in an organization. Companies that do this properly will always outperform, even if initially, it is technically tricky, expensive, and requires a culture change. Companies that persist and successfully navigate the process tend to retain talent, perform better, and promote a corporate culture of accountability.