The Columbo Method: How to Find What's Actually Wrong With a Business

In every episode of Columbo the viewer sees the culprit in the first few minutes. The story is about how the lieutenant proves it. Diagnosing a business works the same way, and the methods are surprisingly similar.

SHORT ANSWER

What the Columbo Method Is

The Columbo method is a way of diagnosing a business by treating the visible symptom as a lead, not a verdict, and then collecting evidence until the real cause is proven. It rests on four habits: noticing numbers that do not match, asking questions that sound naive, looking for evidence outside the official reports, and always coming back with one more question. Most business problems are not hidden. They are misattributed: the owner sees the symptom, blames the nearest suspect, and pays to fix the wrong thing.

Columbo ran on American television from 1968 to 2003, and its structure was unusual. The audience watched the crime at the start, so there was no mystery about who did it. The tension came from watching a rumpled detective in a raincoat notice the one detail that did not fit, apologise, leave the room, and then turn back at the door with his famous line: "Just one more thing."

I first watched it when I was six or seven, at our summer house. When it rained at the weekend, the neighbours invited us in, and one of the channels was showing Columbo. I loved it. Years later, after hundreds of strategic sessions, I realised why it came back to me. In business nobody is murdered, but time, money, resources and opportunities are killed every day, and the investigation works the same way. It is a digital murder of its own kind, and the culprit is rarely the one everybody suspects.

EPISODE ONE

Episode One: The Influencer Budget

A pharmaceutical manufacturer came to us with a business that, on the surface, was working well. It sold sports nutrition and medical products, and a large part of its marketing went through bloggers: specialists in their fields who recommended the products to their audiences. The network was big, around 400 bloggers, and so were the budgets. Sales were good, so nobody was looking for a problem.

The clue was simple. A person contacted the bloggers directly, called them and discussed terms, and the prices they quoted were completely different from the ones the company was paying. It was not one blogger or one mistake. It was systemic: the budgets allocated to bloggers had been inflated two to three times by the department that managed them. At the scale of a factory and a network of four hundred people, that was not a rounding error. The same pattern repeated in paid social: advertising budgets on Instagram and Facebook were set higher than the campaigns needed, and between the company and the actual work sat layers of contractor teams producing content, each adding its own cost.

Nothing in the regular reports pointed to it, because the reports came from the same people who managed the money. The product sold, the campaigns ran, and the losses were hidden inside success.

The ending was practical. The contracts with the bloggers were renegotiated and replaced with clean agreements, and the company kept substantial sums that had been leaking out every month. The money that was freed went into developing its own products and its own websites, which a business with such a complex internal system needed far more than another layer of intermediaries.

We have seen the same story in fitness clubs, where sports bloggers carry much of the promotion. Everyone in a business has their own interests, and that is normal. But the longer some people stay in control of a budget nobody checks, the more they start to treat it as their own, and the same thing happens in sales.

Losses hide best inside a business that looks successful.

We have run hundreds of strategic sessions, and Columbo's line describes most of them better than any consulting framework.

ONE MORE THING

"Just One More Thing": Where the Real Problem Hides

A typical case starts with a confident request. A company wants a new website, although the current one was built on a modern platform only two years ago. The explanation is that the design does not sell.

Then we start asking. When was the move to the new platform? Were the old addresses redirected? It turns out they were not, and two years of search history disappeared on launch day. Or the contact forms send inquiries to a phone number nobody answers and to a mailbox that belonged to a former employee. The design was never the suspect worth arresting. We see this not only in small businesses but in mid-sized projects with serious budgets.

In large companies the culprits are different but just as quiet: too many integrations, one of them failing without anyone noticing, a customer journey and a sales funnel that were configured once and never checked again. What the request describes and what the evidence shows are rarely the same thing, which is why our migration checklist starts with old URLs rather than new designs.

NUMBERS THAT DON'T MATCH

When the Numbers Don't Match

Columbo's best clues were small contradictions: a story that did not fit the timeline, a watch that showed the wrong time. In a business the equivalent is a pair of numbers that should move together and do not.

The most common pair is traffic and conversions. Traffic grows, conversions fall. Behind it there is often a target set from above that nobody dared to question: double the traffic this year. When the goal is badly framed, the execution follows it. The marketing team buys broad advertising, chases queries that are only loosely related to the business, publishes article after article, and the traffic chart goes up while the sales chart goes nowhere.

A few years ago this took a specific form. Companies started producing blog content with prompts as simple as "please write an article about...", and later at scale, a hundred articles at a time pushed straight to the site. For a while it worked and traffic rose. But the result was the same generic text that now exists on tens of thousands of other sites, with no value of its own. Google treats content produced at scale mainly to rank as spam, and we have watched online stores built on this approach lose their traffic as quickly as they gained it. Physics still applies: what was pushed up without weight comes down hard. We wrote about what content is worth when everyone can publish five hundred articles in this article.

Other pairs worth watching:

• many leads, few sales: the problem is often in follow-up, pricing or who handles the inquiry;
• growing sales, flat profit: discounts, returns or delivery costs are eating the margin;
• a busy team, a slow pipeline: people are working around the system instead of in it.

THE SPREADSHEET

The Evidence in Someone's Spreadsheet

Every detective looks for the thing people keep for themselves. In a company that thing is usually a spreadsheet.

Managers do not always like the company's CRM. Some keep their own tables out of habit, some out of carelessness, and some because a private list of clients is easier to take with them when they leave. Wherever people maintain their own file next to the official system, the system is failing at something, and the file shows exactly what. We traced how that happens in Every Enterprise System Started as Someone's Spreadsheet.

In one company where we set up a CRM, an ERP and automation, every sales call was recorded. For calls longer than ten or twelve minutes, an AI assistant produced a condensed summary for the owner, while the manager filled in the CRM as usual. When the two were compared, they described different conversations. The call had been about one thing, and the manager wrote down what suited the manager.

That is why we connect the website, the CRM and the ERP so that the business sees what actually happened, not only what someone reported. Call recording has to follow the consent rules of the country or state where it happens, and the team should know how the records are used. A good setup records inquiries, calls and deals automatically. Instead of a manager's summary, an AI assistant produces a record of what was actually discussed on the call and a short summary of the conversation. The best version is one where the manager cannot change the record, only add comments to it. The evidence then stays intact, whoever is telling the story.

NAIVE QUESTIONS

The Naive Questions

Columbo played the slow, slightly confused man who needed things explained, and people explained far more than they meant to. Simple questions work the same way in a business conversation, because nobody has a prepared answer for them.

These are some of the questions we ask, depending on the conversation:

• What is your main goal, apart from making money?
• What can your company do, other than cause harm?
• Who is your most dangerous competitor, the one nobody has noticed yet?
• Why do people leave your company?
• What would make you fire your best salesperson or your best developer?
• Do you remember what you did last weekend?

None of them is about websites, and every one of them has, at some point, told us more about a business than its marketing report. The last one sounds like small talk. It usually reveals whether the owner still has time to think about the business or only time to run it.

Nobody prepares an answer for a simple question.

THE WRONG SUSPECT

When the Suspect Is the Wrong One

Sometimes the whole room is sure who did it: the website, the advertising, the agency, the market. And the evidence points somewhere else.

When I was learning to drive, my instructor asked me to look at the cars around us and tell him who I thought was behind the wheel. I spotted a Mercedes S-Class and started reasoning: probably a business owner, maybe a senior executive or a government official, judging by the plates. "No," he said. "That's an idiot. He's going to hit you and take someone else with him." He pointed at another car, I started analysing that one too, and he stopped me again: "Another idiot." He was not saying that everyone on the road was a fool. He was teaching me to watch what each car actually does instead of the story I had invented about who was driving it.

Business requires the same discipline. Sales fall and we blame the website, a manager misses the plan and we question their competence, marketing costs rise and we suspect the agency. The website may be fine while nobody follows up the inquiries, the manager may be capable but working with poor leads, and the agency may be executing a brief built on a false assumption. And sometimes the website really is the problem, the manager really is in the wrong role and the agency really is wasting the budget. What matters is not deciding before you have the evidence. Columbo never needed to decide who someone was before investigating what they had done. He watched, listened and noticed what did not fit.

In our experience, roughly seven times out of ten the root of the problem is people. Not bad people, but people in the wrong place: a director, a manager, a marketer or an SEO specialist who is not competent for the role they hold. People who burned out long ago and keep the process at its old speed. People who are hurt by something in the company, and whose resentment turns into quiet sabotage. We say this without wanting to accuse anyone, and with the same care Columbo had: he never raised his voice.

It seems that AI could help here, and for people who want to improve it does. In practice, those who are in the wrong place rarely bother to learn new tools. They are more likely to use AI to craft a better explanation of why the result is not their fault. So far we have never met a vengeful AI that sabotages a project on its own initiative. What comes next, nobody knows, but for now the suspects are still human.

The pattern changes with the size of the business. A company that has outgrown the startup stage and is moving toward mid-size meets it everywhere. In large businesses it shows up in processes: specialists who were excellent twenty years ago have fallen behind the technology, and consciously or not, they slow down everything around them. This is why we do not work only on websites. A site cannot fix a process that someone is holding back, and a new design will not make a funnel work if the person responsible for it has stopped caring.

THE ONLY SUSPECT

When You Are the Only Suspect

The hardest investigation is the one where the owner is also the only witness and the only suspect. A founder who has just started, has little money and does most of the work alone cannot hire a detective, and often cannot afford to be wrong. The same method applies here, turned inward: ask the uncomfortable questions of yourself, not only of an AI assistant, which will politely agree with you unless you push it.

The question from our list that matters most at this stage is the bluntest one: what can you be useful for, other than causing harm? It sounds harsh, but it separates two very different situations. If the product is good, in our experience around 60 per cent of the problem is delegation: the founder holds every task, becomes the bottleneck and has no time left for the work only the founder can do. If the product is weak, good marketing will not save it. It will only make more people aware of a product that does not keep its promise and strengthen the wrong positioning.

Good marketing does not fix a weak product. It amplifies it.

Questions worth asking yourself at this stage:

• Would I buy this at this price if it were someone else's?
• Which three tasks do I do that someone else could do better or cheaper?
• How do customers describe what I do, and does it match how I describe it?
• Where does the money actually come from, and how much of it depends on me personally?
• What am I avoiding because it is uncomfortable: a price increase, a hard conversation, a product decision?
• Am I using AI to think, or to avoid thinking?
• If I stopped working for a month, what would break first?

None of these needs a budget. They need honesty, and that is the one resource an early founder can always afford.

YOUR INVESTIGATION

How to Run Your Own Investigation

An owner can start the investigation before calling anyone:

1. Write down the symptom without a suspect. "Inquiries fell by a third since spring", not "the website stopped working".

2. Find the pair of numbers that do not match, such as traffic and conversions or leads and sales.

3. Check the plumbing before the design: redirects after the last migration, where the forms send inquiries, who receives them and how fast they answer.

4. Look for the private spreadsheets. Every one of them marks a place where the official system fails.

5. Ask one more question of the person closest to the problem, after the meeting has officially ended.

6. Consider the people as well as the tools, calmly and without looking for someone to blame.

Oh, and just one more thing. If you are sure you know who is to blame, check who prepared the report.

If the evidence still points in several directions, an outside view helps. That is what a strategic session is for, and we explain why the problem you came with is rarely the one you have and what changes when you ask whether you are solving the wrong problem.

FAQ

Sources

• Google Search Central: Spam policies, scaled content abuse