What a ten person team should standardise and what not
Three tests for deciding which parts of a ten person business need one agreed way of working and which parts are better left to individual judgement.
The size where everything is still bespoke
At ten people a business usually has two convictions at once. One is that things are getting messy and need structure. The other is that structure is what ruined the last place anybody worked. Both are correct about different parts of the business.
Ten is the size where the founder stops seeing everything. Work now passes between people who were not in the original conversation. Something will break and the question is not whether to standardise but which few things to standardise first, because the capacity to write and enforce standards at this size is roughly one person's spare afternoon a week.
What standardisation costs and what it will not improve
Standardisation does not improve quality by itself. A standard way of doing a thing badly produces consistent bad work and makes it harder to notice. It also will not substitute for hiring properly. A rule exists to remove avoidable variation and not to make an unsuitable person competent.
The costs are writing, enforcement and lost judgement. Writing is the smallest of the three. Enforcement is the real expense because an unenforced standard is worse than no standard at all: the team learns that written rules are decorative, which taxes every future attempt. Lost judgement is subtler. A capable person constrained to the standard path will sometimes take the worse option because it is the approved one.
What a team of this size should leave alone
Do not standardise creative or advisory output. A studio or consultancy selling bespoke thinking damages the product by templating it, and clients paying for judgement can tell when they have received a format instead. Standardise how that work is scoped and invoiced and leave the work itself to the people doing it.
Leave individual working methods alone too. How an experienced person sequences their own day, which notes application they prefer and how they structure their own drafts are not business risks. Nor is the tone of client email, within obvious limits. Imposing a house voice on ten people produces stilted messages and achieves nothing a customer values.
Pricing on complex jobs is the third exception. A rate card works for repeatable work. A one page rule for pricing a difficult job either hands away margin or loses the deal, so keep that as a named person's decision with an approval limit instead.
Three tests for deciding
Ask first whether it touches money or a statutory record. Invoicing, collections, payroll inputs, expense approval and anything feeding your books should have exactly one agreed way. Variation here produces errors that are expensive and slow to find.
Ask second whether it gets handed between people. Client onboarding, job handover, system access granted and removed when somebody joins or leaves. These are interfaces and interfaces need agreement by definition. The work either side of the interface can stay individual, which is the useful principle: standardise the joins and not the interiors.
Ask third whether variation costs the customer. If a client receives a different experience depending on who picked up their request then that inconsistency is visible and it undermines pricing. Response commitments and quote structure belong here. Anything passing all three tests is overdue. Anything failing all three can wait until the business is larger.
How to judge whether it is working
Compliance with the standard flatters. A team can follow a documented process precisely while the process produces the same errors it always did.
What matters is the error rate where work changes hands and the time it takes a new person to become useful. Both improve genuinely when the right things have been standardised. Watch for a third signal in the other direction. If capable people start asking permission for decisions they used to take then the standardisation has gone past the joins and into the interiors, and you will see that in slower delivery before you see it in anybody's complaint.
The common mistake
Standardising the visible things and leaving the money alone. Templates get redesigned, file naming gets a convention and a brand guideline appears. Meanwhile three people each have their own way of raising an invoice and nobody owns collections. This is expensive in the plainest sense. The standardisation effort was spent where errors were cheap and skipped where they are costly, and an unchased receivable is a far larger number than an inconsistent slide.
The other error is adopting a large company's handbook. Policies written for four hundred people assume a compliance function and several layers of management. Dropped into a ten person business they create obligations nobody can meet and everybody learns to ignore.
What changes at the next size step
Standards are sized to the business and should be revisited rather than accumulated. Around twenty five or thirty people the things safely left individual at ten start to need agreement, particularly hiring, performance conversations and anything involving a first layer of managers. Plan to rewrite what you have at that point instead of adding to it. A document that has only ever grown is a reliable sign that nobody has asked recently what it is for.
This guide explains how something works. It is not legal, tax or financial advice, and rules change. Confirm anything regulatory with a qualified professional before you act on it.