Move Your Organization Forward….Even With Limited Data

Unprecedented—“These are unprecedented times.” A word seldom used is now ringing in my ears. From the precall recording at my child’s pediatrician’s office to a text from my Shipt shopper, everyone seems to have gotten on board with the idea that we are indeed in unprecedented times. With the economy in crisis and graphs that show sharp and quick declines, we have to act quickly. How do we move forward though when the data changes are unprecedented? And should we move forward? While the circumstances are more challenging and the analysis more cumbersome, we can still be data-driven decision makers by examining past trends more closely, expanding our data sets, quickly implementing a measurement plan, and finally, acting! We should move forward and that limited data set might just contain the guidance your organization needs.

Look For Your Trend In Historic Data

When we see sudden changes in data that leave us with limited amounts of data, we can expand our data points by identifying if the same change occurred during another period or for just a segment of our data set. It can be helpful to look back at historic data (I find a time period of about 2 years a comfortable place to start). Look to identify if the same trend you are seeing today has occurred in the past. For example, set out to answer the questions—has there been a time when total onsite traffic decreased by 10% for 2 weeks? If the answer is “no,” dig a little deeper. Ask, has there been a segment or a specific channel that has shown the same data trend? 

Now, narrow your date range to look at just that timeframe, and examine how users were engaging in that period. We are most interested in the data that changed during this time. The same traffic sources, goals, and pages will likely be your top performers. However, we want to identify what changed. Given traffic volumes will be volatile, stay close to your rate metrics, i.e., goal conversion rate, percentage of total page views, and percent of traffic. We are interested in pinpointing how user behavior was different during this unique time. Included below are some good questions to ask:

  • Did a specific goal’s conversion rate stay steady while others decreased? 
  • Did a specific article take up a larger percentage of total page views?
  • Did a larger percentage of total traffic come from desktop?

Conduct Strong Market Research

Given our limited data set, we can also increase our indicators by grabbing data from the market. Google Trends is one of my favorite tools to provide an idea of how others are responding to change. Look for industries that haven’t experienced the same volatility, or have quickly recovered. 

Once you have identified a few organizations that have been able to avoid the change you are seeing, put in place a Google Alert to watch. Over a few days, you will gain information about their promotions, sales tactics, and messaging. A pro tip: Track their marketing messages versus their own Google Trend, determining for yourself where they are seeing success. Once you have tracked competitors or other industries, compare this data set with your own. Graph the trends together, pinpointing where similarities and differences are occurring. 

Put In Place a Measurement Plan

Quickly implement a rigorous measurement plan. Your measurement plan should be automated and hone in on the volatile metrics. Dashboards can quickly be created using Tableau, SuperMetrics, or Google Data Studio. Our team prefers the ease of the SuperMetrics Google Sheets plugin. Put a reminder on your calendar to look at the report every morning. Tell stakeholders you will be reporting out. Give yourself accountability—anything to make sure that this task doesn’t get pushed down your to-do list.

Once you have a report in place, now is the time to know your data more intimately. By understanding nuances in your data set, you can learn how users are responding on a daily, or possibly hourly, basis. While the overall trend may lean heavily in one direction, by reviewing small segments of your data, you can identify outliers. Disecting a limited data set by segments provides more opportunity for analysis. Within the larger trend, can you find segments of your traffic that have begun acting positively? If so, what caused the increased engagement? Once you identify the stimulus, it’s simple—do more of that. 

Just Do Something

Given limited data and sudden changes, the stakes are often high and the circumstances uncertain. But, by delaying decision making, you forfeit time that your team could be testing and optimizing. A basic economic principle—there is an opportunity loss for not acting. Discarding a limited data set forfeits your opportunity to learn from the market, study nuances in day to day changes with the data, and lean on past trends seen in historical data. Lean into the guidance that data provides, enabling you to make the most informed decision possible. 

Move forward with confidence!