Price Benchmarking and the Bloomberg Terminal for Insurance

What did Bloomberg actually change?

Picture a bond trader in 1980. To compare rates across government bonds, there was no screen to pull up. You picked up the phone, called around to other traders, did math by hand, and flipped through printed rate sheets that had already gone stale by mid-morning. The data existed and bonds had standard identifiers, so the standardization was there. The problem was access.

That is what makes the story matter. When information is hard to reach, the advantage does not go to the smartest analyst. It goes to whoever has been around the longest. Someone with twenty years on Wall Street and a relationship with every major dealer had an enormous edge, not because they were a better trader, but because they had accumulated knowledge that was not written down anywhere. Experience itself became the moat.

Michael Bloomberg had just been let go from Salomon Brothers in 1981. He was not an outsider guessing; he had worked inside the bond market at the highest level and saw exactly where the friction was. His company, originally Innovative Market Systems and later Bloomberg LP, built the first version of what we now call the Bloomberg Terminal: it collected scattered bond pricing from across Wall Street firms and put prices, analytics, charts, and comparisons on a single screen.

Why is insurance the same story?

Put the pre-Bloomberg bond market next to today's commercial insurance market and the parallels stack up fast.

Start with scale. The US commercial property and casualty market runs around $300 billion a year, putting it in the same weight class as the bond market of the early 1980s. Then look at how deals get done. Neither market has a central marketplace. In bonds, a buyer called a dealer and negotiated a price privately. In insurance, every policy is a private deal between a buyer and a carrier with a broker in the middle, and there is no public record of what comparable coverage costs. The result in both cases is the same: very little price transparency, and a reliance on gut instinct built from having seen enough deals.

The deepest parallel is that the data is physically trapped. Before Bloomberg, Merrill Lynch could see what Merrill Lynch traded, and Salomon could see what Salomon traded, but no single firm could see the whole market at once. To find a real price you called five dealers and pieced together five numbers. Commercial insurance is arguably worse, because pricing data belongs to individual carriers who each see only the policies they wrote, locked across different software systems, wholesalers, and specialty markets.

What does price dispersion mean for a property owner?

Price dispersion is when two essentially identical buildings, same asset class, same geography, same construction type, pay dramatically different amounts for insurance. The hard part is that neither owner knows it. There is no reference price, so when the renewal arrives and the broker says "this is market," there is nothing to check it against.

A broker who has placed a thousand policies in an asset class does develop a feel for what things should cost, but that is pattern recognition built on memory, not a scalable data point, and most buyers do not even have that. They have one broker, one renewal number, and no idea where they sit in the broader market. If you could instead see every comparable transaction, you could finally answer the questions that matter: am I at market, am I getting a good deal, or am I overpaying because my submission happened to land with the wrong broker?

What does this look like inside the Advocate app?

To make it concrete, the hosts brought in David, a product engineer who has been at Advocate roughly four years building the data layer, and who did much of the Bloomberg research behind the series. He pulled up the pricing comps page and walked through a live example.

The flow mirrors Bloomberg's. The pricing already exists, scattered across carriers and brokers, so the work is aggregating and cleaning it, then letting a user filter down to what is relevant: asset class, geography, construction type, distance to coast. Filtering Texas down to Houston with a few attributes turns the full transaction set into a focused, comparable picture showing rate online and premium per door.

He also walked through the newer work: AI-generated reports, and a case-creation feature where you drop documents in, the system structures and organizes them, then benchmarks the policy for price and quality and produces a gap analysis showing how your coverage compares to the market. Underneath is a factor model that isolates which attributes are actually moving price, whether that is distance to coast, construction type, or simply which carrier your submission landed on. As David put it, the crazy part is that without the data, you would never have known the difference.

Why is Advocate's job harder than Bloomberg's?

Here is the key difference. Bloomberg got a running start because the bond market already had standard ways of calculating returns. The foundation was in place, so Bloomberg only had to build the tool on top of it. Insurance has no equivalent. There is no standard way to calculate a rate or a consistent method to compare one policy to another across carriers and states.

So Advocate is building both layers at once: the shared standard underneath and the tool on top. Insurance has never really had either, which is why the hosts framed this as doing the Chicago Board of Trade moment and the Bloomberg moment simultaneously. First you create a shared standard, people adopt it, and the data starts flowing. The platform becomes the place everyone goes for answers, and from there new tools and products get built on the foundation.

Does data put insurance people out of work?

This is the recurring fear whenever technology enters a relationship-driven industry, and the history cuts the other way. When the Chicago Board of Trade standardized grain, it did not shrink the workforce; there are far more people in commodities today than in the 1850s, in roles that could not have existed when the market was too messy to support them. When the Bloomberg Terminal launched, many bond traders worried that shared data would make their experience worthless. Instead the industry grew, new roles appeared, and the traders who embraced the terminal became the most valuable people in the room.