Margin improvement usually starts with subtraction rather than a price rise. Work out the margin you make on each client or service line, decide what to stop doing, price the next new client correctly, and hold the line on scope. Done in that order, most businesses find a meaningful improvement without asking a single existing client to pay more — because the problem is rarely the headline price. It is the work that quietly costs more to deliver than anyone costed.
Margin is the number that decides whether a business is sustainable or merely busy. Revenue can grow while margin shrinks, which is how an owner ends up working harder for less and cannot explain why. This guide is the practical sequence: where margin goes, how to find it in your own numbers, and what to change first.
Where does margin actually go?
It rarely collapses. It erodes, through four routes that all look reasonable at the time:
- Scope creep. Doing more than was agreed without adjusting the price. The single largest cause in most service businesses.
- Underpricing relative to value. Fees set from cost plus a margin, rather than from what the work is worth to the client.
- Unprofitable clients. Busy work that crowds out better work and consumes the capacity you would need to win it.
- Cost drift. Overheads and subscriptions growing without anyone making an explicit decision about them.
Each is invisible month to month and obvious over a year, which is precisely why the fix begins with measurement rather than instinct.
Step one: what is the margin on each client?
Take your client list, or your service lines if you have many small clients, and work out the effective hourly rate or gross margin for each. You need three things: what you billed, the hours or direct costs that went into delivering it, and an honest allowance for the admin the account generates.
Most owners doing this for the first time are surprised in the same direction: the accounts that feel easy are producing far more of the margin than their share of revenue suggests, and the ones that take the most time and generate the most administration are producing the least. Rank the list. The bottom of it is where the next three months of work is.
An invented illustration of the shape — not a client, and not a claimed result. An owner bills £40,000 a month across three types of work. Costed honestly, type A runs at a 42% gross margin, type B at 31% and type C at 9% — and type C takes almost half the hours. Repricing type C and letting the worst of it go drops billing to £35,000 for a quarter, which is where nerve matters. Then the freed hours refill with type A work, and by month nine the business bills the same £40,000 at a blended margin near 36% instead of 29% — roughly £2,800 a month more gross profit, with no extra hires and hours handed back every week.
Step two: what should you stop doing?
Margin improvement is more often about stopping than adding. Three candidates, in order of how quickly they pay back:
- Services that take far longer to deliver than the price supports, and always have.
- Clients who consistently expand the scope and were never repriced for it.
- Internal admin that grew around a problem which no longer exists.
Stopping feels like losing revenue, and for a quarter it usually is. The nerve required is the actual constraint, not the analysis — which is why this work goes better with somebody holding you to the plan when billing dips in month two. It is the most common piece of business coaching there is.
But what if the low-margin work wins you the good work?
This is the honest objection to everything above, and sometimes it is correct. A cheap piece of starter work genuinely does lead to better work with some clients. The problem is that it is also the most comfortable excuse available for not repricing anything, so it needs testing rather than assuming.
Test it with the client list you already have. Take every client who came in through the low-margin service in the last two years and answer two questions for each: did they go on to buy anything at a proper margin, and how long did it take? You are looking for a conversion rate and a timescale, not a feeling. Owners who run this usually find one of two answers. Either a meaningful share converted within a few months, in which case the cheap work is doing a real job and should be priced and capped deliberately as a route in — a fixed scope, a fixed price, and a limit on how many you take at once. Or almost none of them converted, in which case it is not a lead source, it is just cheap work that has been given a strategic-sounding name.
Where it is genuinely a route in, treat it like marketing spend and hold it to the same standard: a known cost per client won, reviewed quarterly, with a ceiling on total capacity it may consume. Where it is not, it belongs on the stop list from step two.
Step three: how do you price the next client right?
You do not need to reprice existing clients to improve the average. Set a target margin for new work, price to it, and only accept work that meets it. Declining work that does not is a legitimate commercial decision rather than a failure, and it gets easier once you can see what the low-margin work has been costing.
Two practical rules. Quote from a standard that includes the admin and the revisions people actually use, not the best-case version. And stop discounting to close — a 10% discount on a job running at a 30% margin removes a third of the profit on it. The Dangers of Discounting model shows that arithmetic in full, and our guide to pricing for profit covers how to build the price itself.
Step four: how do you hold the line on scope?
Scope creep is not usually a client trying it on. It is a series of small, reasonable requests that nobody wrote down. The controls are unglamorous and they work: a written scope for every engagement, a named person who can say "that is outside what we agreed", a simple change note for anything extra, and a quarterly check of logged time against the scope you sold.
When the check shows drift, have the conversation early, with the arithmetic in front of you. Our guide on difficult client conversations covers exactly that conversation, including the retainer that quietly grew from 16 hours to 27.
Step five: what about overheads?
Review overheads deliberately once a quarter rather than in a panic once a year. List every recurring cost, mark each as essential, useful or historic, and cancel the historic ones the same day. Subscriptions, unused licences and services bought for a project that finished are where most of the easy money sits.
Keep it in proportion. Cutting costs alone rarely fixes a margin problem caused by pricing and scope, and cutting the wrong ones damages delivery, which costs more than it saves. If the numbers themselves are the part you avoid, that is what financial coaching is for — regular sessions that build real confidence with your figures, delivered with Buzz Accounting. Our post on why most small businesses plateau covers what happens when this work keeps getting deferred.
What order should you do this in?
- Week one: margin per client or service line, ranked.
- Week two: decide what stops, and what gets repriced.
- Week three: set the target margin for new work and write the standard quote.
- Week four: put the scope controls in place and diarise the quarterly check.
- Then: hold your nerve through the dip, and re-run the ranking in ninety days.
Get the next one in your inbox
One practical, plain-English guide for business owners each week. No spam, unsubscribe any time.
Frequently asked questions
Will I lose clients if I put prices up?
Some, and it is worth knowing in advance which ones you can afford to lose. Rank the client list by margin first, then decide where an increase is justified by what has actually changed: scope that expanded, costs that rose, work that was underpriced from the start. Explain the reasoning in a conversation rather than by email, and give reasonable notice. The clients most likely to leave over a fair, well-explained increase are usually the ones already producing the least margin, which is why the ranking comes first. Losing one of those and refilling the capacity with better work is a good trade, not a loss.
Should I improve margin by cutting costs instead?
Cost control matters, but it rarely fixes a margin problem whose cause is pricing or scope. Cutting an overhead is a one-off saving with a floor; repricing the work and stopping the loss-making services changes the underlying economics and keeps paying. Do the quarterly overhead review because it is quick and it finds real money in unused subscriptions and historic services. Then spend the serious effort on the client ranking. Be careful about cutting anything that touches delivery quality: the saving is visible immediately and the cost arrives three months later as rework, complaints and clients who quietly do not renew.
How do I work out margin per client if I do not track time?
Estimate for one month rather than waiting for a perfect system. Have everyone note roughly how long each job took, in half-hour blocks, for four weeks. That is enough to rank the client list, which is all you need at this stage; the difference between a client at 40% and one at 12% will be obvious long before the data is precise. Add an honest allowance for the admin each account generates, because that is where the real difference between easy and difficult clients shows up. Once the ranking exists you can decide whether ongoing time recording is worth the effort it costs.
How quickly does margin improvement show up?
Expect a dip before the improvement. If you stop or reprice low-margin work, billing usually falls first while the freed capacity has not yet refilled with better work, and that gap is typically one to two quarters. That is the point at which most owners abandon the plan, which is why it helps to write down the expected dip before you start so it does not feel like evidence of failure. Pricing changes on new work show up faster but move the average slowly, because they only affect what you win from that point. Re-run the client ranking at ninety days and compare properly.
How do I end a client relationship without damaging my reputation?
Give the price increase first and let them decide. Most of the clients you want to exit will exit themselves when the work is priced properly, and that is a cleaner ending than a resignation letter. Where you do need to end it, do it in a conversation rather than by email, be honest but brief about the reason — the work has moved away from what you specialise in, or you can no longer deliver it to your own standard at a workable price — and never itemise their faults. Give notice that reflects how dependent they are on you, finish the work in hand properly, and offer one or two names of firms better suited. Handled that way, the client frequently refers people to you afterwards.