One promise, many clinics: writing a franchise operating standard that travels

You standardise across franchise clinics by fixing the promise and freeing the method. Write down the small number of outcomes every site owes the patient, define the piece of evidence each outcome must leave behind, then let each site work out how to produce it with the people and equipment it actually has. Audit the evidence, not the intention. A network that agrees on outcomes and records will look consistent to a patient. A network that agrees only on procedures will not.

That is close to the opposite of how most operating manuals get written. The usual instinct is to describe everything, because everything felt important on the day it was written, and because a thick document looks like control. What you end up with is a four hundred page file that is read once during onboarding and never opened again.

I have written both kinds. The thick one felt safer. It did nothing.

A manual is not a standard

A standard is what happens in the room when nobody is watching. The manual is only a claim about what should happen. The distance between the two is where the actual operating problem lives.

There is decent evidence that the distance is large even among trained professionals with a written protocol in front of them. A 2024 study in the International Journal of Clinical Pharmacy watched community pharmacists in England carry out clinical checks on prescriptions and found they routinely omitted subtasks their own standard operating procedures required. This was not a study of people working under pressure in a chaotic shop. The variation showed up in controlled conditions, which points at professional habit and judgement rather than at stress.

Read that as a design finding rather than a criticism. If qualified people drift from a written procedure while calm, the procedure was never doing the work you thought it was. Something else was.

Separate the promise from the method

Start by sorting everything in the manual into two lists.

The promise is what the patient gets, and it does not vary. The qualified person named on the record performs the procedure. Consent is taken before payment, not alongside it. The price given in writing is the price charged. Every case produces a record the patient can take to their own doctor. A complication has a named owner and a defined response time.

The method is how a site delivers that, and it should vary. Staffing mix, appointment sequencing, which room does what, the local supplier, the language the counselling is delivered in, how the follow-up call gets made. A site in a small town with four staff cannot run the same choreography as a site with fourteen, and forcing it to try produces the worst outcome available, which is a site that fails the standard and hides it.

This distinction matters more in franchising than in owned operations, because you are dealing with capital you do not control. A 2025 systematic review in BMJ Global Health looked at 47 studies covering 29 health social franchises across 25 countries and found the results genuinely mixed. Client volumes and satisfaction improved. Quality and cost-effectiveness did not improve reliably, and the review concluded that outcomes depended heavily on the local environment each franchise operated in rather than on the model itself. The brand does not carry quality on its own. The design of what you fix and what you free is what carries it.

The four line standard

Here is the format I use, and it fits on one page per procedure.

**The promise.** One sentence saying what the patient gets. Written from the patient's side, not the operator's.

**The trigger.** When this applies, and just as importantly when it does not. Most standards are ignored because staff cannot tell whether today's case is one of the ones it covers.

**The proof.** The artefact this leaves behind. A signed form, a photograph set to a defined angle, a timestamped note, a returned call log. If the standard produces no artefact, you have written a preference.

**The local part.** An explicit line saying what the site decides for itself. Naming this is what stops the whole document being treated as advisory.

Monday morning version: take your ten most referenced procedures, rewrite each as those four lines, and archive the rest of the manual as reference rather than as standard. Ten pages that people can hold in their head will beat four hundred that they cannot. If a rule does not survive the rewrite, it was probably a preference wearing a uniform.

Proof beats inspection

The strongest argument for building evidence into the work is that self reported compliance is close to worthless.

A 2025 systematic review and meta analysis in BMC Health Services Research pooled studies of the World Health Organization surgical safety checklist and found overall compliance of 73 per cent, but completeness of only 51 per cent. So the checklist was used in most cases and fully worked through in about half. A 2023 narrative review in the Journal of Thoracic Disease is sharper still. It gathers studies where hospital records showed compliance at or near 100 per cent while direct observation found only a small minority of checklist items were actually being checked, and one where documented time out compliance ran in the mid to high nineties while barely a fifth were correctly performed.

The fix is not more auditors. It is choosing standards whose proof is a by-product of doing the work properly. A photograph taken at a fixed angle cannot be produced without positioning the patient. A consent form timestamped before the payment record cannot be back-dated in a system that locks entries. Build the evidence into the flow and most of your audit stops being a visit and starts being a query.

If your evidence of compliance is a tick from the person who was meant to comply, you have measured willingness to tick.

Let sites adapt, but make them ask

The same narrative review makes a second point worth stealing. Checklists work better when they are adapted to local conditions and existing workflow rather than imposed unchanged, because a step that duplicates something the team already does gets skipped, and skipping becomes normal.

So build a route for variation. A site that cannot meet a method files a short request: what it cannot do, why, what it proposes instead, and how it will still produce the proof. You approve or refuse in writing, with an expiry date. Two things then happen. Deviation becomes visible rather than hidden, and your central team finds out where the standard is unrealistic. The third or fourth identical request is not an exception. It is a drafting error, and the manual is what should change.

What you hold and what you do not

You do not control whether a franchisee cares. You do not control their landlord, their staff turnover, the local salary market, or the day a good manager resigns. You do not control whether the market rewards the standard you hold.

What you control is narrow and it is enough. Which promises are non-negotiable, and how few of them there are. Whether each one produces evidence without anybody being asked. What you do the first time a strong performing site breaks one, because the network reads that decision more carefully than it reads the manual. And whether you rewrite the standard when the exception requests keep telling you the same thing.

Hold that line and the network converges slowly, from the evidence upward. Write four hundred pages instead and you will have a beautiful document, a compliant file, and clinics that do whatever they were doing before you arrived.

Questions people ask

How do you standardise operations across franchise clinics?

Fix the promise, not the method. Write down the small number of outcomes every site must deliver, define the evidence each one leaves behind, and let each site decide how it gets there with the staff and equipment it actually has. Then audit the evidence rather than the paperwork. Manuals describe intentions. Records show behaviour.

Why do franchise operating manuals fail?

Most fail because they are written as complete descriptions rather than as working tools. They grow to hundreds of pages, mix legal requirements with preferences, and get read once during onboarding. A site under pressure falls back on habit. A manual only changes behaviour when it is short enough to remember, specific about what must never vary, and tied to evidence somebody actually checks.

How do you audit franchise clinics without damaging the relationship?

Audit the evidence the work already produces rather than staging an inspection. If every case leaves a record, a photograph set, a consent form and a follow-up note, most of the audit can happen without a visit. Publish the same scorecard to every site so nobody is singled out, share results openly, and reserve visits for the sites the evidence says need help.