Every decade or so reinsurers get fixated on a bit of tech wizardry, and pursue it with vigour. In the 20-teens it was Blockchain. Now it’s AI. Back in the 90s, the internet was the next must-have technological innovation.

Few big reinsurers escaped the temptation to create their very own ‘fast lane’ on the ‘information superhighway’. Everyone wanted an internet strategy, and determined to build their tech from scratch, complete with a custom client portal, bespoke search engine, many-layered directory, a publishing system, and even collaboration tools.

Let me take you back. Swiss Re had a ‘knowledge-management programme’ backed by homemade intranet tech to support information dissemination to clients. It even appointed dedicated ‘knowledge managers’ to ensure the company’s know-how was available to everyone, inside and out.

Shortly after, Munich Re built ‘NATHAN’. The Natural Hazard Assessment Network pushed all the Mighty Munich’s catastrophe data and risk maps out to everyone through a custom-built intranet, with functionality that allowed underwriters to query the hazard anywhere on earth.

In a way these moves were visionary. These reinsurers had a lot of knowledge, and wanted to share it in a reinsurance world where direct relationships were still important, especially in Europe. Digitising the information was revolutionary. It was the component of the information superhighway mania that the sector got right.

Custom building the distribution network, though… that was just daft. By the time reinsurers hade made their own, the emerging generic products were better, faster, and cheaper. The technological plumbing required to publish, search, navigate, authenticate, and distribute information was and remains generic.

Still, big reinsurers and probably thousands of other companies built undifferentiated infrastructure to solve the same problems multiple times: authentication, document indexing, navigation, search, discussion boards, directories, and content management. Most bespoke 1990s code was abandoned with boxy double-breasted jackets, pleated trousers, and tassel loafers.

Flash forward. Today our insistence on instantly available data-about-everything is almost universal. Behind it, though, we have so much data that dissemination is no longer the challenge. The real problem is curation and packaging, and to get that right (using existing tools), reinsurers need to know what information, out of the billions of bytes made daily, is actually useful.

Today’s challenge for reinsurers is to ensure that their cedants and capital providers have the best possible data covering exactly what they need to right now, and nothing much extra.

There’s a tendency, for example, to model a portfolio then send along 300 pages of analytical output.  Most recipients will immediately resort to the “Summarise Long Document” button (perhaps twice or three times). Without advanced agent training, that will skew the main points of the missive in one unknown direction or another, rendering them virtually useless.

Much better to ensure model outputs, risk data, hazard analysis, and other data streams are already trimmed to their pertinent points before hitting ‘send’.

Meanwhile, data floods into insurance companies in ever voluminous cascades. Insures, to varying degrees, use this cavalcade of data effectively, to underwrite well. More than ever, they are able to record and analyse the numbers and words at levels of speed and complexity unimaginable even six months ago.

I can see that the reinsurers’ temptation to collect all that data for their own risk-level analysis is enormous, but that’s like inventing a new internet. It’s been done.

Rather than collecting more data, reinsurers should drop back to the ethos of the 90s, where they sent forth data into the world to ensure that cedants (whether MGAs, syndicates, or insurance  companies) have all the data they need, and that they are using it well. They should think less about the risk data coming up chain, and make sure have two key things:

• the high-quality data they need to underwrite well, and
• the analytical tools to use it wisely, delivering astonishingly low loss ratios (hopefully).

This is much more efficacious than focusing on historical data delivered by their insurers, a rear-view mirror approach that has an important place, but cannot be the only way to look ahead. Current data is key. It should be assessed through a constant deep dive into the data lake of current risk and trends.

To do this, reinsurers must be comfortable and connected:

• Comfortable that the data they receive is of the highest quality
• Comfortable with the way cedants record and analyse data,
• Connected so they can receive data and gain access to outputs.

It’s not enough simply to look at the portfolio in front of you. Re-underwriting primary risks is duplicative. That’s what ceding commission pays for, and if you write excess of loss, you may not even care about the details about the underlying underwriting. You’re concerned about accumulations and how the book will interact with your own portfolio. You need good data to answer those questions, not more.

Gaining access to the better-quality data you need starts with having a very good idea of what data your carriers have. Next, you need to know how they use it. That data and analysis audit needs to be done up front, now, before the capacity is granted. Finally, you can ensure they have the additional data and tools they need to be the best.

That’s what certain capacity providers do when they underwrite MGA binders. They even pay attention to the specific platforms and software being used. If you’re a reinsurer and you don’t do that, take a step back from analysing the froth coming down the information superhighway. Instead, don your mask and flippers, and dive below the surface.