What happens in the first three months?
The honest version, including the part nobody advertises: the first month is mostly checking whether your numbers are dependable, because everything after it is worthless if they are not. Here is the sequence, what arrives when, and how to judge it.
Why the order is this way round
Most founders want the forward-looking work first. The model, the plan, the analysis that answers the decision in front of them. It is the reason they made contact and it is the part that feels like value.
Doing it first is almost always a mistake. A model built on numbers nobody has checked produces confident answers that are wrong, and the wrongness surfaces a quarter later when a decision made on it turns out to have been made on a false picture. That is worse than having had no model, because it was believed.
So the sequence runs backwards from what founders expect: check, then build, then look forward. It takes about ninety days to get through all three, and the third month is where the work starts being worth what it costs.
Month one: find out what is true
The first month is diagnostic. Very little is produced and a great deal is established.
- Understand how the business actually earns. Not the pitch version: which customers, which products, what they pay, what it costs to serve them, and where the money physically arrives.
- Look at the last six to twelve months of accounts against the bank, and find the places where the two disagree.
- Test whether the close is real. Are costs landing in the month they belong to, is revenue recognised on delivery or on invoice, are there provisions missing that an auditor will insist on later.
- Find the metrics already in use and write down what each one currently means, which is frequently the first time anyone has.
- Work out the real cash position: cash less statutory dues already collected and not yet deposited, and less anything committed.
- And ask the team what they do not trust. Somebody always knows which number is wrong.
What you receive at the end of month one is a written assessment: what is reliable, what is not, what is missing, and what that means for the decisions you are currently making. It is short and it is frequently uncomfortable.
Month two: build the base
The second month is construction. The aim is that by the end of it, the numbers arrive on a date and mean the same thing every month.
- A close calendar agreed with your accountant, with a date the books shut and a date the pack goes out.
- The chart of accounts reviewed, so the reporting you want is possible from the ledger rather than assembled by hand each month.
- A reporting pack designed for your business: performance against plan, cash and runway, and the operating metrics that actually decide anything for you rather than a standard set.
- Definitions written down, once, for every metric in that pack.
- A thirteen-week cash forecast, built so your team can update it.
- And whatever provisions were missing put in, so the management numbers stop diverging from what the audit will eventually say.
The first real pack lands at the end of this month. It will be the first time the numbers have been produced this way, so expect to spend an hour arguing about what a line means. That argument is the work, and it happens once.
Month three: start looking forward
With a reliable base, the forward-looking work becomes worth doing.
- Unit economics split by whatever genuinely differs in your business, with the allocation rules written down so it can be reproduced.
- A model built on drivers, with the assumptions separated from the arithmetic, tested against the last six months before anyone uses it to look forward.
- The decision you originally made contact about, now answered on numbers that can be relied on.
- A view of runway that accounts for what growth itself will absorb, rather than cash divided by last month's burn.
- And the second pack, which should take a fraction of the time the first one did.
By the end of the third month there should be a rhythm that continues without anyone deciding to continue it, and at least one decision that was made differently because of what the numbers showed. If neither is true, something has gone wrong and it is worth saying so.
What it looks like when the books are already good
Some companies arrive with a reliable close, a clean ledger and a competent accountant. It is less common than it should be and it changes the shape considerably.
Month one compresses to a fortnight of understanding the business rather than auditing it. Month two becomes designing reporting rather than repairing the base under it. And the forward-looking work starts in week three rather than week nine.
This is worth knowing before an engagement starts, because it changes what is worth paying for. A company with good books does not need the first third of this sequence and should not be charged for it. The assessment in month one is partly there to establish which situation you are actually in.
What you are asked for, week by week
Engagements slow down for one reason more than any other: the information takes three weeks to arrive. It is worth knowing in advance what gets asked for, so it can be gathered before rather than during.
- Week one: read access to the accounting system, the bank, and whatever billing, invoicing or point of sale system the business actually runs on. Plus the last two sets of filed accounts.
- Week one: an hour with whoever knows how the business earns, which is usually a founder, and a second hour with whoever does the bookkeeping.
- Week two: the contracts or pricing for your largest customers, and whatever the sales team uses to track the pipeline.
- Week two: the current headcount with salaries, and any hiring plan that exists even informally.
- Week three: whatever reporting already goes to investors or a board, however rough, and the last shareholders' agreement if there is one.
- Week four onwards: answers to specific questions, usually about why a particular number is what it is. This is the part that needs one person who can chase things inside the company.
None of it needs preparing or tidying first. Sending the messy version is faster and considerably more informative than sending a cleaned-up one, because how the information is actually kept is itself part of what is being assessed.
How to tell whether it is working
Four tests, none of which requires you to understand finance to apply.
- 01Does the pack arrive on the same date each month without anyone chasing it? A rhythm that depends on reminders is not a rhythm.
- 02Can you read it without a translator? If the numbers only make sense when someone talks you through them, the reporting is built for the wrong reader.
- 03Has a decision changed? Not a large one necessarily. A price held, a hire delayed, a customer repriced, a channel stopped. If nothing has changed in three months, the analysis is decorative.
- 04Are you being told things you did not want to hear? An adviser who agrees with you every month is producing comfort rather than judgement.
Say so if any of the four is failing. Most engagements that end badly showed all the signs in month two and nobody raised it until month six.
What is not going to happen in ninety days
- Your margin will not improve because someone looked at it. Analysis identifies where the money goes; changing it is an operating decision your team executes over quarters.
- A fundraise will not be made easier by three months of reporting alone. It helps considerably and it is not sufficient, and anyone claiming otherwise is selling something.
- Your accountant will not be replaced, and should not be.
- Historic problems will not disappear. Provisions that were never made, filings that were late and periods that were closed on incomplete numbers stay in the record. They get named and handled rather than removed.
- And the team will not immediately use the reporting. That takes a couple of quarters and it depends on the pack being genuinely readable, which is why that is a test rather than an assumption.
What ninety days does produce is a base that can be relied on, a rhythm that continues, and a company where the next question takes an afternoon to answer rather than a fortnight. Everything valuable after that depends on those three, which is why they come first.