Three Changes CEOs Must Make for Analytics In the COVID Era
You can’t run a business in 2020 like it’s 2019.
Nearly every CEO would agree with that statement. But businesses are still trying to use their old analytics operating models to cope with a drastically different reality.
Change starts at the top. CEOs must drive their analytics teams to adopt new behaviors, or fall behind. The winners will adjust their tactics now. Some already have. Specifically, CEOs need to direct their analytics teams on the right time frames to analyze, instill urgency as a cultural value around analytics, and adopt a new operating model that prioritizes value from data over perfection.
Why Are People Acting as If Nothing Changed?
Companies are still relying on muscles they used when buyer behavior followed a completely different pattern.
- Airlines are clueless about what ticket prices should look like, and they’re hoping their old models will somehow spit out the answer (that’s not going to happen any time soon).
- Investment models are broken with government bonds and stocks swapping yields for the first time in 70 years.
- In past recessions, luxury item spending dropped, but restaurants and work apparel sales remained fairly consistent. With COVID, restaurants have been forced to slash capacity or risk being the hub of a super-spreader event. More formal apparel sales have plummeted because people don’t see the point in flaunting high fashion when they’re stuck at home. As a result, many retailers have declared Chapter 11, and the restaurant industry saw three times the job losses reported by any other industry.
Historically, management teams have relied on BI for insights. What that really means is whenever the business asked analysts why something’s happening, the analysts relied on their prior experience to decide what to go look at. Then they made guesses about what might be changing behaviors.
Most of the teams in organizations that have invested in analytics automation rely on supervised learning. This form of artificial intelligence uses historical data to predict the future. Unfortunately, that way of looking at the world is completely broken post-COVID. Obviously, historical periods and historical data don’t look anything like what’s happening today, so those techniques don’t work.
People are stubbornly relying on experience to help them predict what will happen next instead of adopting a smarter, more productive way to be data-driven.
Change is hard. In fact, it often needs to be forced. A CEO is in a great position to encourage, or even mandate, change.
Three Things You Can Do Today
Assuming you keep the same teams and automation you invested in prior to COVID, there are still things you can do to correct your course. The world is evolving too fast to do traditional, measured analytics that gets information to the business in months and drives change in years. We need to make 3 changes that drive business impact from data in weeks.
These changes will mean you and the rest of the executive leadership team must work closely with your analytics organization.
The outcomes will be worth the time investment—and you’ll have the data to prove it.
Here’s what you need to do today:
- Give your analytics organization direction on the time frames you want them to focus on
- Change the values of your analytics organization—which means you need to change how you think of analytics too
- Give your analytics team a new operating model
Let’s look at each of these action items in more detail.
Focus On When
There are three periods to be concerned about right now. First, there’s the COVID outbreak. Second, there’s a period of transition when the country is reopening. And finally, there’s the post-COVID “new normal.”
I thought these phases would be distinctive, but I’ve been forced to adjust my expectations based on the data. One could argue whichever stage (or muddy in-between) you’re standing in depends on where you and your customers are located.
Some states elected to ignore COVID CDC guidelines and have remained open without mandating any changes in behavior. Some are experiencing a second lockdown after reopening and experiencing a resurgence in cases. Others have stayed in some form of lockdown since March and aren’t planning on relaxing any time soon.
You need to determine whether you want:
- Your team to try to figure out buying behaviors today. This would help an organization struggling with sales pivot their strategy immediately.
- Them focusing on what’s likely to happen in the near term. If the governor of a state has announced a major change, what’s likely to happen next? Will buying behavior rebound, remain stagnant, or rebound and then drop if another wave of new cases hits?
- The team to figure out what new behaviors are likely to last when things go back to “normal.” This helps form a long term strategy that the rest of the organization can rally around.
A Shift in Values
This is very important for CEOs to focus on because it involves changing their own behavior too.
Up until this point, the analytics team has focused on establishing confidence. Every result that they showed you went through weeks or months validation. They looked at the data from a bunch of different angles and made sure it was perfect before risking a presentation. They want to feel a high level of confidence in everything they do.
Waiting for a perfect result can’t happen right now.
You need the analytics organization to investigate data and come up with insights on what’s happening on a daily basis. You need them sitting down with business leaders to talk about results fast and frequently.
Every day is ideal because things are changing that frequently.
The business leaders will be empowered with knowledge about what’s going on this minute and the analytics team receives feedback on what (and when) the business wants them focused on next.
Today’s values should not be wrapped up in high confidence. Analytics organizations should embrace exposing findings quickly so the business can react intelligently.
The Right Operating Model
The current model most organizations use creates a bottleneck from analysis to insights, loading much of the human work at the beginning where data preparation slows everyone down while ultimately culling much of the valuable data out of the analysis.
The model needed today focuses the AI on data preparation insight discovery and aiding in the prioritization of which insights to act on. This speeds up the analysis and brings the business leads the ability to act quickly to changes that are happening in the business now. It involves tracking the impact of the change on the business. It then guides the business team on how to react to their new information.
Ultimately, this model ensures you measure what’s happening through every step of the funnel. This allows you to see how they impact revenue because, at the end of the day, everything needs to be tracked in terms of ROI.
An Even Better Way to Do Things Exists
The data models that still work today are unsupervised learning models. This is because unsupervised learning is meant for people who aren’t sure what’s happening. The machines help them figure out what has changed by elevating insights and then digging into what business leaders feel is most relevant.
Unsupervised learning doesn’t require someone with experience to tell it where to look and what to look for. Data selection bias is eliminated. It also has the power to analyze unstructured data and find previously unattainable insights.
These techniques work very well right now, but many organizations don’t have unsupervised learning deployed yet.
In the meantime, you need to accept that the world has changed and broken existing techniques, then help your analytics team change how they work. Your team needs your guidance to focus on the right time, shift their values, and use the right operating model that bases everything on ROI.