Legal Spend
When should we use an AFA on litigation?
Updated July 2026
Use an AFA where the matter type is high volume and predictable enough that historical cost data lets you price a fixed or blended fee with confidence: routine auto and trucking defense, employment, and similar repeat work. Avoid it on low-volume, high-variability, or novel matters, where neither you nor the firm can estimate scope, and the fee becomes a gamble for whoever guessed wrong.
What makes a matter type a good fit for an AFA?
Three conditions together: enough volume that you have handled many similar files, enough predictability that those files cost and run within a tight range, and enough historical data to prove it. When all three hold, a fixed or blended fee sits close to real cost, so neither side is exposed. Miss any one and the fee is priced on hope, not evidence.
| Matter profile | Fee fit | Why |
|---|---|---|
| High volume, low variability (routine auto, employment) | Strong AFA candidate | Deep closed-file data supports a fixed or blended fee close to real cost |
| Moderate volume, some variability | Fixed fee per phase | Fix a price on the predictable phases, keep the uncertain ones hourly |
| Low volume or high variability | Keep hourly | Too few comparable files to price a fair fixed fee |
| Novel or first-of-a-kind | Keep hourly | Neither side can scope the work, so any fixed fee is a guess |
Which matters should stay on hourly billing?
Low-volume, high-variability, and novel matters. If you handle only a handful of a certain case type a year, you lack the data to price a fair fixed fee, and so does the firm. Complex or first-of-a-kind litigation can balloon in ways no one can scope up front. On those, hourly billing keeps risk where it belongs instead of forcing a guess into a fixed number.
- Low-volume matter types: too few comparable files to build a stable cost range.
- High-variability types: cost swings so widely that any single fee misprices most files.
- Novel or bet-the-company litigation: scope cannot be estimated in advance by anyone.
- Matters where the firm itself cannot commit to a scope: a shared blind spot, not a deal.
What is the decision framework in practice?
Run each matter type through four questions in order: Do we have volume? Is cost predictable within a range? Do we have the historical data to prove the range? Can the firm scope the work too? Four yeses point to a fixed or blended fee. A no on data or predictability points to hourly, or to a per-phase fee that only fixes the predictable stages.
- Volume: have we handled enough of this matter type to know how it behaves?
- Predictability: do those files cost and run within a tight, repeatable range?
- Data: do we have the closed-file history to prove that range, not just a sense of it?
- Firm scope: can the firm estimate the work with the same confidence we can?
How do you phase into AFAs on a book you have never priced?
Start with your highest-volume, most routine matter type, because that is where your data is deepest and the fee is least risky to set. Prove the pricing on that category, refine it against actual results, then extend to the next most predictable type. Keep novel and low-volume work hourly until you have enough closed files to price it. Expand from evidence, not ambition.
This sequencing matters because a single mispriced AFA on the wrong matter type can sour a firm relationship and set the whole program back. Beginning where your data is strongest lets you build both the pricing discipline and the firm's trust before you extend the model to harder categories.
Common questions
What is the single biggest factor in deciding whether to use an AFA?
Predictability, backed by data. If you have handled enough of a matter type to know what it costs and how long it runs, and you have the closed-file history to prove that range, an AFA is a reasonable structure. If the matter type is rare, highly variable, or genuinely novel, no fee estimate can be trusted, and hourly billing protects both sides from a bad guess. Volume and predictability travel together: the matter types you see most often are usually the ones that behave consistently enough to price. Start your AFA program there and leave the unpredictable work on hourly until your data catches up.
Outside counsel spend by matter type→Can we use an AFA on complex or high-stakes litigation?
Rarely for the whole matter, and never without real caution. Complex, high-stakes, or first-of-a-kind litigation can move in directions neither you nor the firm can scope in advance, so a flat fee either exposes the firm to an unbounded loss or exposes you to an overpayment built into the price. Where you want some cost certainty on a complex matter, a fixed fee per phase on the predictable early stages, with the uncertain stages left hourly, is safer than a single flat fee for the entire case. The rule holds: fix a fee only on the parts of the work your data can actually price.
Flat fee vs hourly for insurance defense→How many closed files do we need before pricing an AFA?
Enough that the cost and cycle-time range for that matter type is stable rather than swinging with each new file. There is no universal number, because a tightly clustered matter type stabilizes faster than a variable one. The practical test is whether adding more closed files still moves your average and your spread meaningfully. If it does, you do not yet have enough history to price a fee with confidence. If the distribution holds steady as files are added, you can price against it. This is why high-volume matter types are the right place to begin: they reach a stable, priceable distribution far sooner than low-volume work.
Do AFAs actually save money?→Should the firm agree to an AFA it cannot scope either?
A firm that cannot estimate the work should not be pricing a fixed fee any more than you should. When both sides lack the data, an AFA simply transfers risk to whoever negotiated worse, and that damages the relationship regardless of which way it breaks. The healthiest AFAs are ones where both you and the firm can look at the same cost history and agree the number is fair. If the firm genuinely cannot scope the matter type, that is a signal to keep it hourly until enough files close to make the economics visible to both of you.
Selecting and assigning defense counsel→CaseGlide is the litigation intelligence platform for Fortune 500 legal departments and insurance claims organizations. It structures live litigation data from defense counsel into executive decisions: reducing defense spend, settling the right cases sooner, and shrinking litigated claim volume.
Next step · See it on your docket
See what your litigation portfolio is telling you
A 30 minute walkthrough on your own docket. No slides, no committee.