Verify then Trust

Verify → then Trust

Several years ago, I recorded a video about something I believed was pretty important.

Trust but verify.

It’s an old Russian proverb made famous by Ronald Reagan during the Cold War.

And at the time, I thought it was pretty good advice.

Actually…

I still do.

I’m just not sure I have the order right anymore.

Trust but verify

The video was about companies working with outside agencies and service providers.

I’d heard several stories from business leaders who had hired capable firms, spent a bunch of money, and ended up disappointed.

The natural conclusion was:

They hired the wrong people.

Sometimes they probably did.

But I wasn’t convinced that explained everything.

These companies had evaluated firms, reviewed proposals, checked references, and chosen people they believed could do the work.

Then they signed the contract.

And waited for the magic to happen.

They trusted.

The problem was that they often weren’t particularly clear about what they were trusting them to do.

What specifically were they trying to accomplish?

What would good performance look like?

How would they know whether things were working?

That was where I landed at the time.

Trust your partners.

But verify their performance.

I still believe that.

But something about Trust but Verify has started bothering me.

The sequence.

Maybe we have it backwards

Verification doesn’t mean distrust.

Quite the opposite.

We shouldn’t begin every relationship, analysis, or decision assuming something is wrong.

Skepticism and suspicion aren’t the same thing.

Verification is simply asking whether something deserves the confidence we’re placing in it.

Think about someone you’ve worked with for years whose judgment you trust.

You probably didn’t trust them on Day 1 the way you trust them today.

Something happened in between.

They made commitments.

They delivered.

Their conclusions held up.

When they were uncertain, they told you.

When they were wrong, they acknowledged it.

Over time, something accumulated.

Evidence.

Eventually you stopped checking everything.

Not because verification became less important.

Because verification had already taught you something.

This person can be trusted.

Which suggests something different:

Trust isn’t the alternative to verification.

Trust is what verification earns.

We have to trust

Organizations couldn’t function otherwise.

We trust dashboards.

We trust Finance.

We trust analysts.

We trust processes.

We trust systems.

Trust allows us to move without independently validating everything every time.

The problem isn’t trust.

The problem is unearned trust.

A dashboard isn’t trustworthy because it’s a dashboard.

A model isn’t trustworthy because a data scientist built it.

A number isn’t trustworthy because Finance provided it.

And an answer isn’t trustworthy because an AI delivers it confidently in beautifully constructed sentences.

Those things may absolutely deserve our trust.

Verification helps us know when they do.

Verification isn’t proof

Verification doesn’t mean proving something is true.

In most business decisions, that’s an impossibly high standard.

The question is whether the evidence is sufficiently credible for the decision we’re asking it to support.

A decision involving a few thousand dollars may require one level of verification.

A decision involving $50 million probably deserves another.

A directional analysis may be perfectly adequate for one decision and woefully inadequate for another.

Verification isn’t about eliminating uncertainty.

It’s about determining whether what we’re relying on has earned enough credibility to carry the weight we’re placing on it.

And when it has?

Trust it.

Verify → then Trust

There’s danger at the other extreme too.

If every conclusion must continually be challenged, every analysis rebuilt, and every person required to prove themselves from scratch, we haven’t created decision integrity.

We’ve created paralysis.

Trust is valuable precisely because it allows us to stop verifying everything all the time.

Good processes earn trust.

Good work earns trust.

Good judgment earns trust.

Good evidence earns trust.

And that trust compounds.

That’s why I’ve come to prefer a slightly different formulation.

Not:

Trust but Verify.

Instead:

Verify → then Trust.

For lack of a conclusion…

Think about something important your organization trusts.

A metric.

A model.

A process.

A partner.

A person.

Maybe even an AI.

Then ask:

Why do we trust it?

Because we’ve always used it?

Because somebody important told us to?

Because it usually looks about right?

Or because, when subjected to scrutiny, it has repeatedly held up?

Trust is incredibly valuable.

But trust shouldn’t simply be assumed.

Trust should be earned.

Verify.

Then trust.