Who is an intermediary's real customer?

On 13 January 2023 Tesla cut its car prices by up to 20 percent. The Model Y got 13,000 dollars cheaper overnight. The real shock landed on the used-car market: according to Edmunds data, the average listing price of a 2020-or-newer used Tesla fell from 76,000 to 58,000 dollars within seventeen days, and time on the lot stretched from 24 days to 39. Used-car platforms such as Vroom had to rebuild their inventory strategy. Tesla made a decision. The sellers adjusted.

Some companies set the product and the price. Others have to live with what was set. I sort business relationships along that line. In a healthy one, both sides sit at the table and the decision comes out of the conversation. In an unhealthy one, one side speaks and the other side learns the outcome.

The invisible asymmetry in insurance distribution

That asymmetry has been carrying the seeds of a quiet crisis in insurance distribution for a long time. Quiet, because while the market is good it shows no symptom: carriers open budgets for new customers, the intermediary grows, everyone is content.

To see the size of the dependence, read the intermediaries’ own reports. MediaAlpha, one of the largest digital insurance intermediaries in the US, states in its 2025 annual report that its largest customer accounts for 25 percent of revenue and the next largest for 24 percent. Two carriers, half the company’s revenue. The top twenty customers make up 82 percent, up from 72 percent a year earlier. The concentration is rising, not falling.

The tap turns both ways, and that is what the structure means in practice. In 2022 carriers cut their advertising budgets; GEICO alone cut 38 percent, roughly 800 million dollars, and digital intermediaries’ revenues collapsed in those quarters. In 2024 the tap opened again: Progressive alone spent 3.5 billion dollars on advertising, up 187 percent on the previous year, and the same intermediaries reported record growth. Two extremes in two years. The intermediary decided neither.

From outside these read as budget news. The asymmetry is surfacing.

The unwritten rule is this: if a B2B company’s revenue depends on one or two customers, that company is not an independent actor but an extension of those customers. It believes it has a strategy of its own. It lives inside someone else’s.

Insurance distribution adds a twist. The intermediary often misidentifies who its customer is. It sees itself as the consumer’s adviser; the party that sets its income is the carrier. Its behaviour follows the money, without anyone saying so.

The same picture in a market like Turkey, only quieter

I can guess the first objection from colleagues at home: this is an American problem, distribution there runs on advertising budgets. The agent-carrier relationship in our market is different.

Partly true. Here the intermediary’s income does not hang on ad budgets, but it hangs on commission, and changing the form of a dependence does not change its nature. In the field I see this: most agents write the bulk of their premium with two or three carriers. The scene always plays the same way. One morning a circular arrives from the carrier: commission in this line will be this much from this date. There is no signature under it and no discussion. The agent does the arithmetic. If that carrier is forty percent of the portfolio, the cost of objecting is forty percent of the year’s income. The agent does not object. The agent adjusts. Commission is set unilaterally and changed unilaterally. Tesla’s price and GEICO’s ad budget worked the same way.

Concentration on this scale carries a cost nobody puts on a spreadsheet. The agent drifts toward recommending the product of the carrier that holds most of its book, even when the better cover for the customer sits with another company. Nobody does this out of bad faith. Protecting the source of your income is the most natural reflex there is. But at that moment the intermediary stops being the customer’s adviser and becomes the carrier’s sales arm, without noticing the change.

AI speeds up the equation

A consumer looking for car insurance used to type a search, click the first comparison site and fill in a form. Today the same consumer asks an AI assistant; in the US, inside ChatGPT, you can get live prices in eleven states and buy the policy right there. I covered that in The customer doesn’t know what they don’t understand. The work an intermediary long earned its keep on, gathering information and lining prices up side by side, is now done elsewhere, cheaper and more tailored to the individual. The consumer’s route no longer has to pass through an intermediary. If it does, there needs to be a reason.

On the broking side I have watched this for years. A broker’s real capital is independence: the ability to work with several carriers and search, without a thumb on the scale, for the right answer for the client. Once that independence goes and the work shrinks to carrying one carrier’s quotes, the broker turns into that carrier’s remote sales arm. The difference is invisible from outside. It shows up in a crisis.

Independent value comes from one of three places: data nobody else can reach with ease (the customer’s claims history, renewal behaviour, real risk profile), expertise that cannot be copied (asking the right question on a non-standard risk), or a real relationship of trust with the customer (running the claim on their behalf when the bad day comes). None of these is built overnight. Once built, neither a carrier’s budget nor an AI assistant can switch them off with ease. As information gets cheaper, judgement and trust get more expensive.

What we did

Balancing the asymmetry does not come from diversifying. In a concentrated market like ours, a small agency working with ten carriers on equal terms is not realistic. The way through is to own the customer relationship and the customer’s data.

We built our own CRM for that reason. A carrier’s system knows the customer as a policy number, and each carrier sees only its own policy. On our screen the customer sits as a whole: which cover they hold with which carrier, what happened in their claim three years ago, what they pushed back on at the last renewal, which cover they have never used. At renewal we look at our own record; we build the offer around what that record says, not around the carrier’s campaign of the month.

As long as that data is ours, which carrier we place the risk with is a secondary question. When commission changes, we are not alone at the table, because the customer stays with us, and the carrier knows it. An intermediary who knows its own customer stays strong at the table whichever carrier it works with. The question is not who you work with. It is who the relationship belongs to.

Where I stand

A company’s most expensive dependence is the one it has stopped noticing because it has grown used to it.

Showing the price fastest is worth less every year; AI does that now too. The intermediary of the future will be the one that produces value independent of both the carrier and the consumer. That value can be real expertise, a relationship of trust, or data nobody can copy with ease. Our answer was data and the relationship; someone else’s answer may be different. For the intermediary with no answer, the question is plain: was the party we took for our customer ever our customer at all?