A few years ago, I worked in a new business in a public company. At the end of the first year, I worked on the annual review with the cofounder. Everything seemed great until the financial forecast showed the new business would not generate the results as the management expected from the very beginning. It turned out before the annual review, nobody had put all the business data in one place and interpreted what it meant. In the subsequent year, the company made a restructuring in the new business based on the insights from the annual review.
Later on, I worked in several other companies in various industries, and bad decisions similar to the above one happened. Most of the time, they were outcomes of one-sided data.
Many companies have realized data analysis can be a superpower for decision-making. They hire talents and install data collecting software to implement data-driven strategies. However, few companies have the setup to understand big pictures with data.
Data from all dimensions need to be taken into consideration to understand business performance and potential improvements. However, when teams set goals and metrics independently, and no one organizes to connect the metrics, it's unclear how each team's work affects one another.
For example, marketing pursues to drive traffic to the company, product tries hard to engage users on their platform, and sales work to land and expand as many customers as possible. Every company dreams to understand which customer cohort converts the most, where they come from, and how to judge if changing a less-than-ideal cohort's behavior is possible and how. No one team can answer these questions alone.
Besides, many companies use metrics before testing their relevance to profit. They hypothesize what metrics may affect business, yet they never take the step to verify the hypotheses before using them in OKR.
What's worse, many companies have data scientists running from request to request. These arrangements drain data scientists' brainpower, take away their time to think strategically, and stop them from bringing real business value to the company.
And lastly, many analytics software is dragging companies away from connecting the dots. Without significant investment, it's very challenging for companies to make third-party software talk to each other and provide a big picture of businesses.
The first step is to gather data team and domain experts to map out business processes together. The data system should be a "digital twin" of the business processes. To achieve the goal, teams need to understand what to track on each node. A simplified B2C product user journey is illustrated below.
After mapping the user journey, the next step is to ask the right questions. Example business questions are listed in the chart above. According to the questions, the third step is to brainstorm related metrics and form hypotheses. For example, to understand where customers come from, teams may rely on click-on ads, impression, and page visit metrics.
Next, teams need to collect data and verify hypotheses with visualization or statistical methods. It’s critical to test metrics before using them in OKR. Otherwise, metrics may lead to biased investments like market focus, pricing strategy, to name a few.
When different stakeholders agree upon metrics, teams can now build a dashboard to monitor how business decisions affect metrics and profits. Sankey diagram is a great way to show how customers move along the pipeline.
When companies create their data strategies, they often start by considering what tools they need before mapping their customer journey and forming hypotheses. As shown in Chart 1, a comprehensive data strategy requires a sophisticated data stack to provide different capabilities. Since the data analytics ecosystem is still fragmented, companies may need to deal with more than ten vendors to build a data stack. If companies start by evaluating vendors for each capability, they may be drowned in information quickly and lose the big picture. Therefore, companies are advised to plan their data strategies with a top-down approach.
Finally, companies should review metrics regularly. Metrics relevant in one scenario may be irrelevant in another. When businesses evolve, so should metrics.
To learn more about the topic from a hands-on perspective, you can refer to Supercharge your monetization - A practical way to lead scoring. If you want to discuss data strategy more in depth, feel free to reach out to Ivy. To stay on top of business strategy and analytics, please subscribe to my newsletter. See you in the next article.