COVID-19 Economic Recovery & Metrics

When I work with a new CEO client, I do two things. First, I ask them to explain the business to me in qualitative terms — Who is the customer? What problem do you solve for them? Who are the competitors? etc.

Second, I ask for specific metrics from their business. This is how you can tell what’s working and not working in the business.

This is how you solve problems in many fields.

I’ll give you two examples from the medical field.

First, let’s look at the COVID-19 pandemic response here in the United States.

There’s a big debate going on right now about when and how to reopen the economy. Do we optimize to save lives or save livelihoods?

For me, the far bigger issue is that we don’t have the data to make a well-informed decision.

Imagine for a minute that we had the ability to test every person in the United States to see if they have COVID-19. Imagine everyone takes a 5-minute test every 48 hours for two weeks.

In that time frame, we would be able to identify exactly who is contagious and who is not. With that accurate data, we could separate the contagious from the not contagious.

Thirty days later, the pandemic would be over in the United States. We would be able to save the maximum number of lives and livelihoods.

(To keep it that way, we would give the same series of tests for anyone entering the United States from another country.)

Notice how when you have really good data, you can make extremely effective decisions.

Let me give you another example.

As some of you know, by “day,” I write and teach my students on CaseInterview.com about how to have a successful career. By “night,” I volunteer as an emergency worker in the Seattle metro area.

When I did my basic life support medical training, I learned how to take baseline vital signs, memorized what ranges were considered normal, and learned how to interpret specific patterns of deviations from the norm.

During training, I was quizzed on scenarios like this:

If a patient’s heartbeat is 120 beats per minute while at rest, their breathing rate is 30 breaths per minute and their blood oxygen level is 85%, what does that mean?

It turns out that vital signs tell a story of what’s going on with a patient.

My job when doing a patient assessment is to capture the initial data set, track those metrics over time, and figure out the story the data is telling me.

In the case of the patient above, their heart is racing (when it shouldn’t be), they’re breathing too fast (as if they’re running, even though the patient is at rest), and the oxygen level in the blood is dangerously low.

Once I understand the story, it informs my decision about what to do next.

In this case, the patient needs supplemental oxygen and immediate transport to a hospital for a higher level of care.

As I started to get better at understanding, recognizing, and interpreting these key metrics, I had an epiphany.

It occurred to me that looking at metrics and finding the story that the numbers are telling is something I’ve done ever since my first job working at McKinsey.

When I read financial statements, I don’t see a page of numbers. I see a story.

I see a unicorn (a company with a $1+ billion valuation) that’s growing fast that can never mathematically be profitable.

I see a company with failing sales that’s actually two businesses commingled — one that’s failing and one that’s actually doing fine.

I see a sales and marketing process that’s over-invested in sales and under-invested in marketing.

Every financial statement and key performance indicator tells a story.

Understanding the story that the numbers tell informs you of what to do next.

Every problem has a corresponding solution. The question is whether you see the problem correctly or not.

If you don’t track metrics or can’t interpret them, you risk misunderstanding the problem. If you misunderstand the problem, you can’t solve it effectively.

When I look at the COVID-19 data set here in the United States, my consistent reaction for the past two months has been that we have a major data-quality problem.

We haven’t done random sample testing to eliminate selection bias in case counts.

We haven’t done random sample testing by geography to quantify the case count by region.

It is very difficult to solve a problem when we can’t even quantify the magnitude of the problem.

Is this a little problem or a big problem?
Is it a big problem or an enormous problem?
Is the problem in the North or the South?
Is the problem in rural areas or urban areas?

When the honest answer is “nobody really knows for sure,” that’s a telltale sign of a data problem.

When I ask a client, “Who are your top ten customers?” and they don’t know… that’s a data problem.

When I ask a client, “What are your contribution margins by product line?” and they don’t know… that’s a data problem.

When I ask a client, “Which customer segments have the longest lifetime value?” and they don’t know… that’s a data problem.

When you’re faced with a data problem, the solution isn’t to just blindly guess at what to do next. The solution is to solve the data problem.

You never want to fly an airplane with defective instruments.

You never want to fire a gun with your eyes closed and “hope” you hit your target.

You never want to run a business by guessing what to do next.

Here’s the obvious yet profound insight.

You solve data problems by recognizing that you have a data problem then actually solving the data problem.

When you ignore the data problem and hope you get it right anyway, there’s a name for that…

It’s called luck.

Don’t get me wrong. I’m absolutely in favor of benefiting from good luck.

At the same time, I prefer to rely on competence.

When it comes to executive-level decision making, competence means making data-driven decisions.