Selling Into Insurance Claims: The Four Numbers a Claims Leader Actually Manages To

Claims leaders don't buy software

I have listened to a lot of reps pitch insurance carriers, and the pattern is almost always the same. The rep opens with the product. Automation. AI. A platform. Something about "transforming the claims experience." And the claims VP on the other end says the thing claims VPs say, which is "send me some information," and that is the end of it.

Here is what I think is going on. A VP of Claims is not compensated on having good software. They are compensated on moving a small set of operating numbers in the right direction, and their bonus, their credibility with the CFO, and their ability to defend headcount all sit on those numbers. When you pitch a capability, you are asking them to do the translation work themselves — to figure out where your thing lands in their P&L. They will not do it. Not because they are lazy, but because six other vendors called them this month and they have a pending count to worry about.

So the job of a seller going into claims is to walk in already attached to one of the numbers they manage to. There are four that matter. Cycle time, loss adjustment expense, leakage, and adjuster attrition. If your pitch does not land on one of those four, you are a nice-to-have, and nice-to-haves get deferred to next year's planning cycle forever.

The four numbers

Cycle time

Cycle time is days. From first notice of loss to closure, usually, though most claims orgs track a family of interim clocks too: FNOL to first contact with the insured, first contact to inspection or assignment, inspection to first payment. A claims leader almost never has one cycle time number. They have a grid — by line of business, by severity band, by whether the file went to litigation.

This matters because "we reduce cycle time" is a meaningless sentence to them. Reducing cycle time on a low-severity auto physical damage claim is a completely different problem from reducing it on a commercial property claim with a coverage question and a public adjuster on the other side. The first is a throughput and routing problem. The second is a documentation and expert-scheduling problem. If you say the phrase without naming the segment, you have told them you do not know the business.

Cycle time is also the number most tied to things the claims org gets publicly beaten up about — customer satisfaction scores, complaint volume to the state department of insurance, agent escalations. So it is emotionally live in a way the cost numbers are not. That makes it a good number to open on and a bad number to build the whole business case on, because it is hard to convert days into dollars in a way a CFO will sign.

Loss adjustment expense

LAE is what it costs to handle claims, as distinct from what the claims themselves cost. It splits into allocated LAE — the money you can tie to a specific file, like defense counsel, independent adjusters, engineers, appraisers, experts — and unallocated LAE, which is the claims organization itself. Salaries, offices, systems, management.

This is the number your CFO-adjacent economic buyer cares about, because it rolls into the expense ratio and from there into the combined ratio. When a carrier is running hot on combined ratio, LAE is one of the few levers they can pull inside a single planning year, because you cannot re-underwrite a book overnight and you cannot reprice in every state at once.

If your product reduces the number of files that need an IA deployment, or shortens the hours a desk adjuster spends assembling a file, or takes work out of the litigation track, you are an LAE story. Say so in those words. "ALAE" and "ULAE" are not jargon to a claims leader, they are the vocabulary of their monthly review. Using them correctly buys you more credibility than any slide.

Leakage

Leakage is money paid out that should not have been paid out, under the terms of the policy and the facts of the file. Overpayment against policy limits or sublimits. Depreciation not applied. Deductible not taken. Subrogation opportunity identified too late or not at all. Duplicate payments. Coverage that should have been questioned and was not. Fraud indicators that nobody worked.

Carriers measure it through claim file audits and quality review programs, which means the number is an estimate produced by a small internal team reading a sample of files. That has two consequences for you. First, the number exists, so a claims leader can usually tell you their leakage picture if they trust you. Second, the number is soft, so any promise you make against it will be scrutinized by someone whose job is auditing files for a living.

Leakage is the highest-value story of the four, because it hits indemnity dollars rather than expense dollars, and indemnity is the bigger pool. It is also the one where overclaiming will end you fastest.

Adjuster attrition

The fourth number is people. Adjuster turnover, time to license a new adjuster in the states you write in, ramp time to full caseload, pending count per adjuster, overtime spend, open requisitions that have been open too long.

A lot of sellers treat this as an HR problem and skip it. It is not an HR problem. It is the number underneath the other three. A short-tenured adjuster closes files more slowly, leaks more, and generates more re-work and more supervisor time. When a claims leader cannot hire, everything else on their scorecard degrades at once, and they know it. If your product takes low-judgment work off an adjuster's desk, or shortens the time a new hire needs before they can carry a real caseload, that is an attrition story and it is often the one that gets you a second meeting.

Which number does your product actually touch

Pick one. Reps want to claim all four because all four sound good, and the effect of claiming all four is that a claims leader believes none of them.

The honest test is this: if your product worked perfectly for a year, which number would move first, and who inside the carrier would notice without being told? If the answer is "the adjusters would like their day more," you have an attrition and productivity story, and you should build the business case on ULAE and ramp time, not on cycle time. If the answer is "more subrogation referrals would get made," that is leakage, and you should be talking to whoever owns recovery. If the answer is "files would move through triage faster," that is cycle time, and you need to name the segment.

Secondary numbers are fine. Lead with one, mention the second as an effect, and let them ask about the rest.

Getting a claims VP to tell you their actual figure

None of this works if you are guessing at their numbers. The whole point of naming the four is to get them to hand you one of theirs, because a business case built on their figure is a business case they have already partly written.

They will not just tell you. Claims leaders are careful people by training — the job is literally about not saying more than the file supports. So you have to ask in a way that is easy to answer and does not feel like a data grab. A few patterns that work.

Ask about direction before magnitude. "Is auto cycle time better or worse than it was this time last year?" is a question anybody will answer. Once they answer it, the follow-up — "what's it sitting at now?" — is a much smaller ask than it would have been cold.

Ask about the segment they are unhappy with. "Where's your cycle time grid ugliest right now — is it the low-severity stuff piling up or the complex files aging out?" This gets you the number and the emotion attached to it, and it signals that you know cycle time is not one number.

Ask about what the audit found. For leakage, "when your quality review team read files last quarter, what were the top two or three things they kept finding?" is far more productive than asking for a leakage percentage. They will tell you about missed subrogation or depreciation errors, and now you are talking about a specific failure mode rather than an abstraction.

Ask about staffing in terms of the desk. "What's a pending count per adjuster look like on the property side right now, and where do you want it?" The gap between those two figures is your attrition story in one sentence.

Ask about what they had to explain. "What's the number your CFO asked you about most recently?" I love this one. It tells you which of the four is actually in play this planning cycle, which is usually not the one you assumed.

The sequencing of all this — how far into a first call you can push before it feels like an interrogation — is the whole subject of the twenty-five minute discovery structure I use for claims conversations, and it is worth planning before you dial rather than improvising. And if you are still trying to earn the first conversation, the difference between an opener that names a claims number and one that names your product is the difference between a call and a voicemail, which is why the insurance cold call script opens on their operating reality rather than yours.

Carriers and TPAs are different businesses

Reps lump them together and then wonder why the deal behaves strangely.

A carrier pays the losses. Leakage is their money, LAE is their expense line, and the claims org is a cost center defended against the CFO every budget cycle. Their buying motion is slow, committee-driven, and heavily gated by procurement, vendor risk, and legal.

A third-party administrator does not pay the losses. They handle claims for someone else — a self-insured employer, a captive, a fronting carrier, an MGA program — and they get paid a fee, often per claim or as a share of premium. That flips the economics. For a TPA, cycle time and LAE are their own margin, so efficiency arguments land fast and land with the person who owns P&L. Leakage, though, is the client's money. A TPA cares about leakage mostly as a competitive story they can tell in an RFP or a client stewardship review, and as a way to avoid losing an account.

That difference changes your pitch. To a carrier, leakage reduction is a return-on-investment argument. To a TPA, leakage reduction is a retention and new-business argument, and you should be asking about their upcoming client renewals and RFPs, not their combined ratio.

Two other structural things about TPAs. They usually handle to each client's specific handling instructions, so anything you sell has to be configurable per account, and "we do it one way" is disqualifying. And they are sometimes required to work inside the client's system of record, which means your integration story has to survive contact with a stack they do not control.

Claims and IT run on different clocks

This is where good deals die quietly.

The claims organization runs on an operational rhythm. Monthly pending reviews, quarterly scorecards, catastrophe season, year-end close. When a claims leader gets excited about your product, they are thinking about the next two quarters.

IT runs on a multi-year rhythm, and in most carriers that rhythm is dominated by a core system program — a Guidewire or Duck Creek or Sapiens implementation, a data platform migration, a mainframe decommission. Those programs consume the integration capacity, the architects, and the political oxygen. A new vendor showing up mid-program is not evaluated on merit, it is evaluated on whether it adds risk to the program.

On top of that you have vendor risk review, security questionnaires, and — if there is any model in your product — model governance and increasingly a regulatory conversation about how decisions are explained. And there are freeze windows. Many property carriers will not push changes during peak catastrophe season, and almost nobody wants a change during year-end close.

What to do about it: ask early and directly what core system they are on and where they are in that roadmap, and ask what their change freeze calendar looks like. Then build your timeline around it instead of pretending it does not exist. A rep who says "I know you're mid-Guidewire, so I'd rather scope this as a desk-level pilot that doesn't touch the core" sounds like someone who has done this before. A rep who forecasts a Q4 close through a change freeze sounds like someone who has not.

Pilots in claims get judged on adoption, not accuracy

The last thing reps consistently get wrong. You will build a pilot, and you will assume the pilot succeeds if the output is good. Accurate estimates. Correct triage decisions. Well-identified subrogation referrals.

That is not what gets measured. What gets measured is whether adjusters used it.

Adjusters live inside the claim system of record all day and they are measured on pending count and closures. Anything that lives in a second window, requires a second login, or adds a step before the thing they were already going to do will be abandoned in about a week, no matter how good the output is. And when the pilot review happens, the sponsor will pull the share of eligible claims where the tool was actually touched, and if that number is low, nobody will care that the accuracy was excellent on the files where it was used.

So design the pilot around usage. Pick one unit or one line, not a scattered sample across the org. Get the frontline supervisor bought in, because they are the person who either reinforces the new step in a huddle or quietly tells the desk not to bother. Ask for the tool to sit where the adjuster already is. And agree the success criteria in writing before you start, in their language — usage rate, files touched, a specific movement on one of the four numbers within the pilot window.

The demo is where this starts, because carrier buyers do not watch demos to be impressed, they watch to find where the thing breaks on their weird files. That is why running a claims demo means inviting the hard cases rather than avoiding them, and why the adjuster in the back of the room is the person you are really presenting to. Get that right and the renewal conversation eighteen months later is a formality instead of a rescue mission.

What I would do next

If I were selling into claims tomorrow, I would write down the one number my product moves, then write the five questions that would get a claims VP to tell me their figure on that number, then say those questions out loud until they sound like curiosity instead of a survey. That last part is the part everybody skips, and it is the part that decides whether the call goes anywhere. We built DrillCall so reps can run those reps against a claims buyer who pushes back the way a real one does, before it costs them a live meeting.

The four numbers are not a framework. They are just what the person on the other end of the phone gets paid on. Show up already talking about one of them and you stop being a vendor call and start being a conversation.