Contingency practice has a specific financial shape that makes ordinary forecasting useless. Revenue arrives in irregular lumps you do not control, while payroll and rent arrive on the first of every month with total reliability. A firm can be genuinely healthy on paper and still have a bad March.
The forecast that works is built from your case pipeline, not from your revenue history. Your pipeline is knowable. Your revenue history is a record of when lumps happened to land.
The model
Four inputs per stage of your case lifecycle.
How many matters are currently in that stage. This comes from your case management system.
What percentage of matters that reach that stage eventually resolve with a fee. This comes from your closed case history, and it is the number most firms have never calculated.
The average fee on those resolutions, which you should segment if your case types differ materially.
The average time from that stage to resolution.
Multiply matters by resolution rate by average fee and you get expected revenue from that stage. Spread it across months using the time-in-stage figure and you have a cash forecast that reflects the cases you actually hold.
Then subtract your fixed monthly burn, which you already know, and read across the row. The month where the line goes negative is the month you need to be planning for right now.
Why the historical numbers are the hard part
Every input above comes from your closed case data, and most firms cannot produce it cleanly.
The usual problems: stage transitions were never recorded with dates, so time-in-stage is unknowable. Closed cases do not distinguish between resolved with a fee and closed without one, so resolution rate is unavailable. Fee amounts sit in accounting, unlinked to the matter record in case management.
Fixing that is unglamorous and it is the whole project. If your system does not stamp a date when a matter changes stage, start doing that today, because in six months you will have the data and today you do not.
In the meantime, reconstruct what you can from the last two years of closed files. It will be approximate. An approximate forecast built from your real pipeline beats a precise average of past revenue.
Where AI does the work
This is the kind of modeling that would otherwise mean hiring someone or paying a consultant for a spreadsheet you cannot modify.
Give it your closed case data and ask for resolution rates and time-in-stage by stage and case type. It will do the analysis and, more usefully, tell you which segments have too few cases to be meaningful. That caveat is one an eager consultant sometimes skips.
Then have it build the forecast model. If you build it as a working tool rather than a spreadsheet, it can pull live pipeline data and update itself, which is the approach in build vs buy and it is what I run.
Scenario testing is where it earns its place. What happens if signings drop 30 percent for a quarter. What happens if the average resolution time stretches by two months, which is the scenario that actually kills contingency firms. What can we afford to spend on marketing while keeping three months of runway. Those questions take seconds to answer once the model exists and are nearly impossible to answer by intuition.
The numbers on the wall
| Number | Cadence | Why it matters |
|---|---|---|
| Cash on hand and months of runway | Weekly | The only number that can end the firm |
| Case inventory by stage | Weekly | Your future revenue, in its current form |
| New signings | Daily | Leading indicator for everything downstream |
| Fees collected | Weekly | Actual cash, against the forecast |
| Cost per signed case | Weekly | Whether marketing is working |
| Fixed monthly burn | Monthly | The line the forecast has to clear |
These belong alongside the operational metrics on the firm dashboard rather than in a separate financial report, because the decisions they drive are the same decisions.
What to keep with your accountant
Tax planning, entity structure, compliance filings, and anything touching trust accounting. Modeling is not accounting, and a forecast is not a tax position. The forecast tells you what is coming so you can decide. Your accountant keeps you compliant and should see the model rather than replace it.
Do this today
Pull your closed cases from the last two years and calculate one number: the percentage that resolved with a fee, out of everything you signed.
That single figure is the foundation of the whole model, and most firm owners are surprised by it in one direction or the other. Either way, you now know something about your business that you were previously estimating.