Is starting a small marketing agency worth it in the UK with two clients already signed?

Executive Summary & Verdict

Promising — 7/10. Start it. The reason this scores where it does has nothing to do with the market for marketing services, which is crowded and always has been. It is that you are starting with £63,600 of contracted annual fee income and roughly £13,000 of year-one costs, which means the business is profitable before you have sold anything new. Almost no agency gets to begin from there, and the ones that do not are the ones that fail.

The £18,000 is doing a different job than you probably think. It is not funding losses. It is funding the gap between doing the work in month one and being paid for it around day 45 to day 60, plus the front-loaded setup spend. On those numbers it is adequate — with one exception, which is the single most important finding in this report and is set out in section 06.

Three things keep this off a higher score. First, concentration: £3,500 of £5,300 a month is one client, and a founding client who signed on a relationship is not a client who has tested you against the market yet. Second, capacity: at nine delivery days a month for two accounts you have about three days of genuine slack, and the third retainer does not fit. Third, the ad spend: agreeing to fund £45,000 a month of platform spend against an £18,000 buffer and 45-day payment terms is not a cash-flow inconvenience, it is the mechanism by which small agencies with good clients go under. Section 12 explains both the money and the VAT consequence, and section 07 explains how to take that risk off the table in a single conversation before you incorporate.

Idea Validation

The idea is already validated, and it is worth being clear about what that does and does not mean. Two clients signing before you trade proves that two specific people will buy specific work from you at a specific price. That is a much stronger starting position than a positioning statement and a Notion doc. It does not yet prove that a stranger will.

The specialism is the part that is genuinely well chosen. "Paid social and lifecycle email for UK DTC outdoor and cycling brands" is narrow enough that a brand founder can tell in one sentence whether you are for them, and narrow enough that your case studies compound instead of cancelling out. Generalist agencies compete on price and proximity; specialists compete on having seen the same problem forty times. Eleven years in one vertical is a real asset and most new agencies do not have one.

What is unproven is the middle of the funnel. Your two clients came from your address book, which is a finite resource — most founders discover it holds two or three deals, not ten. The question the first six months must answer is whether inbound from the specialism produces a third client who has never met you. Until it does, you have a well-paid two-client consultancy rather than an agency, and that distinction decides whether year two grows or plateaus.

Market & Target Customer

Your market is not "UK DTC". It is the specific band of brands that can afford a £3,000-plus monthly retainer but cannot justify a £55,000 in-house hire, in one product category, in one country. That band is small, and being able to name it is an advantage rather than a limitation.

What defines a client you can actually serve profitably

• Revenue between roughly £1m and £5m. Below that, the ad budget is too small for paid social to move anything and the retainer eats the margin. Above it, the brand starts hiring a performance lead and you become a supplier to that person rather than the strategist.
• An owner who still makes the marketing decision. Your entire value is senior judgement delivered directly. The moment there is a marketing manager in the middle, your day rate becomes an execution rate.
• Already spending on ads, badly. A brand with no paid history is a longer, riskier engagement than a brand with eighteen months of messy account structure you can fix in week two.
• An email list they have never properly worked. Lifecycle email is where a specialist shows measurable value fastest, and it is the half of your offer with no media cost attached to it.
• Seasonality you understand. Outdoor and cycling brands peak in spring and at Christmas. A vertical specialist can plan around that; it also means your own revenue is correlated across clients, which section 15 returns to.

The uncomfortable part of a narrow vertical is that the number of qualifying UK brands is probably in the low hundreds, and a meaningful share already have an agency. That is survivable for a one-person business that needs six to eight clients at maturity. It would not be survivable for a fifteen-person agency, which is worth knowing now rather than in year three.

Competitive Landscape

You are competing in three directions at once and they have nothing in common with each other.

Against full-service agencies, you win on seniority and lose on breadth. Published UK pricing puts full-service retainers at £3,500–£16,750 a month [1], and at the bottom of that range a brand is buying a team in which the person who sold the work is not the person doing it. Your pitch is that at £3,500 they get the senior person for all of it — a strong argument, and a promise you personally have to keep every month.

Against freelancers, you win on nothing structural. A senior freelance paid-media specialist charges £70–£100 an hour or £490–£700 a day [2] and can do a version of your work. What separates you is holding both channels together, a retained commitment rather than a booked week, and enough process that a client is not exposed when you take a holiday. If you cannot articulate that difference, you are a freelancer with a company number, and you should price accordingly.

Against the client doing it in-house is the competitor most agency plans forget. Every brand in your band is one good hire away from not needing you, and the trigger is usually the moment your retainer approaches the cost of a salary. At £3,500 a month you are £42,000 a year, which is close enough to a mid-level in-house salary that the comparison will get made out loud. The defence is not a lower price. It is being demonstrably better than the person they could hire for that money, and saying so with numbers.

Business Model & Monetisation

You are selling your own time on a retainer, and until you hire, that is the whole model. The numbers behind it deserve to be written down explicitly because almost every agency founder gets a version of this wrong in year two.

Start with available days. Roughly 260 weekdays, less bank holidays and twenty days of leave, leaves about 232 working days, or 19 a month. Selling, finance, admin, your own marketing and the client conversations that are not delivery will take 35–40% of that. Call it twelve delivery days a month. Your two accounts consume about nine of those — six for the paid social and email client, three for the email-only client — which leaves three days of slack for everything unplanned.

Now the price. £5,300 a month across nine delivery days is an effective rate of £589 a day. UK agency day rates for a mid-to-senior practitioner are published at £600–£1,500 [1], and a senior freelancer's own day rate runs £490–£700 [2]. So your pricing is defensible, and it is at the bottom of the defensible range. That matters for one reason: it leaves no room to buy someone else's day.

Where the revenue can come from, in order of how well it fits this model

• Monthly retainers. The core. Predictable, and the only line that makes the business saleable later.
• Paid, fixed-fee audits as a front door. A £1,200 two-week account and lifecycle audit converts strangers into clients better than any proposal, and gets paid either way.
• Lifecycle build projects. A one-off flow build or migration is scoped, bounded and priced at a project fee rather than a day rate. It is the natural upsell inside an existing retainer.
• Percentage of ad spend. Common in the industry and wrong for you. It ties your income to a number you are supposed to be optimising, and it drags you into the invoicing arrangement section 12 tells you to avoid.
• Reselling freelance days. Viable, but only at the right markup — see section 10. At your current prices it is not viable at all.

Finances & Budget

This is a cheap business to start and an easy one to get into trouble in, and the two facts are related. Year-one contracted fee income is £63,600 (£5,300 × 12). Year-one non-labour cost, itemised below, is under £13,000. Those two retainers sit in the middle of the published UK band for agencies serving £500k–£2m brands, £1,500–£3,000 a month, and at the low end of the band for £2m–£5m brands, £2,500–£5,000 [2]. There is no premises, no stock and no equipment beyond a laptop you already own.

That leaves £5,053 of the £18,000 uncommitted, against roughly two months before the first invoice clears. It is tight but it works, because £5,300 arrives every month thereafter and £63,600 less £12,947 leaves £50,653 available for your own salary and dividends. As a first year from a spare room, that is a good outcome, and it is the reason this report says start it.

One more number worth holding in your head. Your break-even, including a modest £2,400 a month of personal drawings, is around £3,500 a month of fee income — which is exactly one client. That is the whole argument for the concentration risk in section 15: you are one cancellation away from break-even, not from failure. Knowing which of those two it is changes how calmly you can negotiate.

If you do hire, the arithmetic changes sharply. Employer's Class 1 National Insurance is 15% above a secondary threshold of £5,000 a year, the Employment Allowance is £10,500, and the national living wage for those aged 21 and over is £12.71 an hour from April 2026 [3]. A £38,000 salary therefore costs you meaningfully more than £38,000, and it needs about £8,000 a month of new fee income behind it before it is safe. That is two and a half new clients, not one.

MVP & What to Build First

No pivot is needed here — the plan is sound. What it needs is a sequence, because a one-person agency with nine delivery days already committed can only build one thing at a time, and doing them in the wrong order is how the good version of this business turns into a treadmill.

Build first, before you incorporate

• The ad-spend conversation. Ask both clients to hold their own platform billing. Ten minutes, no cost, removes the largest risk in the plan. Do this before you sign anything, because it is a much harder conversation in month five.
• The master services agreement and the processor terms. You will be handling their customer lists and their ad audiences, which makes you a processor and makes a written Article 28 contract a legal requirement, not a nicety [1]. One solicitor-drafted set of terms you reuse for every client is the single highest-value £1,500 you will spend.
• Payment terms you can live with. Retainers billed in advance on 14 days, not in arrears on 45. Your larger client has proposed 45 — that is negotiable, and it is worth more to you than another £200 on the fee.

Build in the first ninety days

• One repeatable reporting artefact. A monthly one-page view of spend, revenue, contribution and lifecycle performance. Built once, it saves a day a month per client and becomes your sales collateral.
• A documented onboarding. Access, tracking, list hygiene, naming conventions, baseline metrics. Client three should cost you two days to onboard, not five.
• A £1,200 fixed-fee audit offer. Your front door for people who have never met you, and the only piece of new-business machinery you have time to build this year.

Defer until there is a third client

• Any hire, of any kind. See the arithmetic in section 06.
• A brand refresh, a proper website, a podcast, a newsletter with a schedule. A one-page site that says what you do, for whom, with two case studies and a way to book a call is enough — none of the rest sells the fourth retainer as fast as the third case study does.
• Adding a third channel. Paid social plus email is the specialism. Saying yes to SEO because a client asked is how specialists become generalists with worse margins.
• VAT registration is the exception to "defer" — see section 12, where registering voluntarily and early is probably the right call.

Go-to-Market & Launch Plan

You have two clients and no pipeline. That is the actual launch problem, and it is not solved by announcing the agency.

• Get both founding clients to a result you can publish, fast. A named client with a specific number is the only asset that sells a third retainer. Agree in week one what metric you will be judged on and ask, in writing, for permission to write it up at 90 days.
• Work the partner channel, not the audience. In DTC the referrals come from Shopify agencies, 3PLs, email platform partner managers and fractional CFOs who serve the same brands and do not compete with you. Five good relationships there outperform any content programme available to a one-person business.
• Go where the vertical physically is. Trade shows and industry events for outdoor and cycling brands put you in a room with your entire market, and almost no agency in your category bothers to attend as a specialist rather than a sponsor.
• Do not launch with a discount. A "founding client rate" sets your ceiling for two years and tells the market what you think you are worth. Your effective day rate is already at the bottom of the published range [1].
• Say no to the first out-of-scope enquiry. There will be one in the first month — a website, a bit of SEO, a brand project. Taking it costs you three delivery days and a year of positioning.

Marketing & Growth

An agency that markets itself badly is a specific kind of embarrassing, and an agency that spends its delivery days marketing itself is a specific kind of insolvent. With three slack days a month you get to pick one channel and do it properly.

Pick the one that compounds with delivery. Writing up what you actually did last month for a client — the account restructure, the flow that lifted revenue per recipient, the creative test that failed and why — costs you an hour because you did the work anyway. Public teardowns of brands you do not work with have the same property and double as your audit lead generator. A content programme requiring original research does not, and will be abandoned by March.

Referrals need a mechanism, not goodwill. Ask every client, at the 90-day review, for one introduction to a brand founder they respect. Put it in the review agenda so it happens without you having to summon the nerve. This is the highest-converting channel a specialist agency has and almost nobody systematises it.

Growth is a price question before it is a volume question. Going from £63,600 to £90,000 does not require three more clients; it requires raising the two you have at renewal and adding one. Published retainer bands put a mid-market agency programme at £3,000–£8,000 a month [1] and full-service work well above that. You are billing £3,500 for two channels delivered by a senior practitioner. The room is there.

Pricing Strategy

Your prices are fine for a founder who delivers everything personally and wrong the moment anyone else touches the work. That transition is the pricing decision that actually matters, and it is worth doing the sum now rather than discovering it under pressure.

Take the £3,500 client at six delivery days a month. Hand that to a mid-weight freelance paid-social specialist — call it £500 a day, which sits between the mid-level freelance band of £280–£490 and the agency mid-weight band of £600–£850 [1][2] — and you pay out £3,000 to keep £500. A gross margin of 14%, with you still carrying the client relationship, the quality risk and the invoice. For that same scope to survive being delivered by someone else at a sane 60% gross margin, the retainer has to be around £7,500. That is inside the published mid-market band [2], which means the number is achievable — but it is more than double what you are charging.

How to price so the business can grow

• Price the outcome and the scope, never the hours. Hourly pricing caps you at your own capacity and punishes you for getting faster at the thing you have done for eleven years.
• Set a floor and hold it. A minimum retainer — £2,500 is defensible on these numbers — protects the twelve delivery days from being eaten by small accounts.
• Quote fees separately from media, always. Published retainer figures are routinely stated excluding VAT and excluding ad spend [3], and your invoices should be just as unambiguous. This is a commercial habit with a tax consequence, covered in section 12.
• Build the annual increase into the contract, and charge for the audit. A stated 12-month review removes a conversation you would otherwise avoid, and a £1,200 audit attracts people with a budget rather than people who want free work.

Legal Setup & IP

Incorporating is cheap and fast. A private limited company costs £100 to register digitally with Companies House, or £124 by post [1], and after that you must file a confirmation statement at least once every twelve months — £50 digitally, £110 on paper, with up to 14 days' grace after the review period ends [2]. Miss it and Companies House may issue a financial penalty and may strike the company off the register [2]. Put both dates in a calendar the day you incorporate.

The limited company is the right structure here, and not mainly for tax. You will be signing master services agreements that carry indemnities, processing other people's customer data, and spending other people's advertising budgets. Doing that as a sole trader puts your house behind every clause.

Engaging freelancers: the part agencies get wrong

The moment you buy a freelancer's day you are in employment status territory, and there are two separate questions. The first is employment law status — employee, worker, or genuinely self-employed — which determines rights, and where the categories in UK law are set out by HMRC and Acas [5]. A long-running, exclusive, you-control-the-hours arrangement can make someone a worker with holiday-pay entitlement whatever the contract says.

The second is off-payroll working (IR35), and here the small-company position works in your favour. Where the client is a medium or large private-sector organisation, the client determines the worker's status. Where the client is small, that responsibility sits with the worker's own intermediary instead [3]. An organisation is medium or large if it meets two or more of: turnover over £10.2m, balance sheet total over £5.1m, or more than 50 employees [4]. You will not be close to that, so a freelancer working through their own limited company assesses their own status — but you must confirm your size if the worker or the party you contract with asks [4], and the exemption applies only to the private and voluntary sectors. It also does not follow you as you grow. Diarise a check when you pass fifty people or £10.2m, which is a nice problem to have.

Everything else to settle before the first invoice

• The master services agreement. Scope, notice period, payment terms, late-payment interest, an IP assignment that transfers on payment rather than on delivery, and a cap on your liability. That cap is the most valuable clause in the document.
• The Article 28 processor agreement. A written contract is mandatory when you process personal data on a client's behalf, and it must specify the subject matter, duration, nature and purpose of processing, the type of personal data and categories of data subject, plus the eight compulsory terms — documented instructions only, confidentiality, security, prior authorisation for sub-processors, assistance with data subject rights, breach and DPIA support, deletion or return at the end, and submitting to audits [6]. Section 14 explains why this clause decides who pays when something goes wrong.
• Advertising code responsibility. Under CAP Code rule 1.8 primary responsibility for observing the Code falls on marketers, but agencies, publishers and other service suppliers involved in preparing or publishing marketing communications also accept obligations under it [7]. Writing a client's claim does not make it the client's problem alone.
• Sub-processor flow-down. If you bring in a freelancer who touches client data, they need a back-to-back agreement and your client needs to have authorised them. You stay liable for their compliance [6].
• Name and IP. Check the trade mark register before you print anything, secure the domain and handles together, and keep your own methodology, templates and reporting assets explicitly excluded from the client IP assignment. Your process is the only asset the business owns.

Tax, Invoicing & Admin

This section contains the one piece of arithmetic most likely to cost you real money, so it comes first. How you invoice client ad spend decides your VAT position.

Compulsory VAT registration is triggered when taxable turnover exceeds £90,000 in any rolling twelve months; you then have 30 days from the end of the month you went over to register, and your effective date is the first day of the second month after you crossed [1]. Taxable turnover is "the total value of everything you sell that is not VAT exempt or 'out of scope'" [2] — which is where the £45,000 a month becomes dangerous.

HMRC distinguishes a recharged cost from a disbursement. A cost your business incurs itself in supplying its own services is not a disbursement, and is part of your supply. VAT Notice 700 puts it plainly: the consideration for a supply includes any payment given to cover your costs in making that supply, unless you incur the costs as an agent [4]. To treat a payment as a disbursement outside the scope of VAT, HMRC requires all of the following to apply [3]:

• You paid the supplier acting as your customer's agent.
• Your customer received and benefited from the goods or services.
• Paying was the customer's responsibility, not yours.
• You had the customer's authority to pay.
• The customer knew the goods or services came from another supplier, not from you.
• The cost is shown separately on your invoice.
• You pass on the exact amount, with no markup.
• The goods or services are additional to your own services.

Meeting all eight for platform ad spend is possible but genuinely difficult, and "paying was the customer's responsibility, not yours" is hard to sustain when the platform contract is in your name. If HMRC treats the spend as part of your supply, your taxable turnover is £603,600 rather than £63,600 — mandatory VAT registration, 20% output tax to account for on more than half a million pounds of media, and no margin in any of it. Do not decide this yourself. Put the exact wording of both client arrangements in front of an accountant before you issue the first invoice. The clean answer, and the one section 06 recommends, is that the client pays the platforms directly so the question never arises.

The rest of the admin, in the order it matters

• Register for VAT voluntarily, early. Your clients are VAT-registered businesses, so charging them VAT costs them nothing, while you recover input VAT on software, the accountant and freelancers. On fee income of £63,600 you are under the £90,000 threshold [1] and can choose — and the choice is probably yes. Once registered, digital record-keeping and software-filed returns apply; confirm the current requirements with your accountant.
• Corporation tax. The small profits rate is 19% on profits up to £50,000, the main rate is 25% above £250,000, and marginal relief applies in between; both thresholds are reduced pro rata for short accounting periods and by the number of associated companies [5]. On roughly £50,000 of pre-salary profit you are at or near the small profits rate, which is one of the better arguments for the company structure.
• Director payroll. Employer's Class 1 NI is 15% above a £5,000 secondary threshold and the Employment Allowance is £10,500 [6]. The salary-versus-dividend split for a single-director company is a real optimisation and an annual conversation with your accountant, not a one-time decision.
• Invoice in advance, chase on day one. Retainers billed on the 1st, due on the 14th, with a stated right to interest on late payment. The habit is worth more than the clause.
• Keep the ad-spend receipts separate from everything else, even if the client is paying the platforms. When the accountant asks, you need to be able to show that money never touched your accounts.
• Companies House deadlines. Confirmation statement £50 digital, annual accounts, and a corporation tax return [7]. None of it is difficult; all of it carries a penalty.

Tech, Hosting & Security

Your technology stack is small and your security exposure is not, because the thing you hold is other companies' customer lists and other companies' ad accounts. A one-person agency is a soft target holding valuable credentials, and clients in your band are starting to ask about it in procurement.

• Never take ownership of a client's ad or email account. They own the asset, you get admin access. This protects them at renewal and protects you from inheriting their billing liability — the same principle as section 06's ad-spend recommendation, applied to logins.
• A password manager with per-client vaults, and hardware-backed two-factor on everything. An ad account takeover on your watch is an incident you report to your client and possibly to the ICO. Treat SMS codes as inadequate for accounts that can spend money.
• Do not hold client customer data anywhere except the client's own platform. The exported CSV sitting in your downloads folder is the most likely breach in this business by a wide margin. Work in their instance, export nothing, and if you must, delete it the same day.
• Write down your sub-processor list. Your reporting tool, your cloud storage, your project management — each one that touches client personal data is a sub-processor your client has to have authorised [1]. Maintaining that list is also how you answer a security questionnaire in twenty minutes instead of a day.
• Cyber insurance alongside professional indemnity. Budgeted in section 06. Read what the policy actually requires of you before you rely on it.

Regional Compliance & Privacy

This is the section a new agency skips and then regrets, because three separate regimes apply to you at once and none of them cares that you are one person in a spare room.

1. UK GDPR — you are a processor, and that is a position with a price

Running ads and lifecycle email for a client means handling their customers' personal data on their instructions. That makes the client the controller and you the processor, and the ICO's guidance on the distinction is explicit that the same organisation can be a controller for some processing and a processor for other processing [1] — you are a controller of your own prospect and employee data and a processor of your clients'.

The distinction decides liability. A written Article 28 contract is mandatory and must set out the subject matter, duration, nature and purpose of the processing, the type of personal data and the categories of data subject, together with eight compulsory terms [2]. The clause that should focus your attention is this one: if a processor acts outside the controller's instructions as to the purpose and means of processing, it will be considered a controller in respect of that processing and will have the same liability as a controller [2]. In plain terms — the day you decide on your own initiative to upload a client's customer list to build a lookalike audience, or to bring a new tool into the workflow, you may have stopped being a processor and started being a controller with full exposure. The protection is not diligence, it is documented instructions.

You also have to pay the ICO data protection fee. The tiers are £52 for a micro organisation (turnover up to £632,000 or no more than ten staff), £78 for small and medium, and £3,763 for large, with an automatic £5 discount for paying by direct debit [3]. Sole traders are included. You are tier 1; pay it, because an unregistered agency handling client customer data is a bad look in a procurement conversation long before it is an enforcement problem.

2. PECR — the rules that govern the email half of your business

Under PECR you must not send electronic mail marketing to individuals unless they have specifically consented to electronic mail from you. The soft opt-in is a narrow exception for existing customers: you may market similar products and services without prior consent only where you collected the contact details yourself, gave a clear chance to opt out when you collected them, and include an opt-out in every message. It applies to past customers, not to prospects and not to purchased lists [4]. You must also not disguise or conceal your identity and must give a valid contact address for unsubscribing, and act on opt-outs promptly [4]. Corporate subscribers — companies, partnerships, LLPs and public bodies — are treated differently and can be emailed without prior consent, though the ICO recommends honouring objections and keeping a do-not-email list as good practice [4].

Two practical consequences for an agency. First, the consent belongs to the client, not to you — an audience the client acquired cannot be marketed to on behalf of a different client, and "we sent it from their domain" is not a defence. Second, the stakes changed recently: the Data (Use and Access) Act 2025 raised the maximum PECR penalty from £500,000 to UK GDPR levels of £17.5m or 4% of global annual turnover, with the increase taking effect on 5 February 2026 [5]. Insist on seeing how a client's list was built before you send to it, and refuse the send if the answer is vague. That refusal is the most valuable thing you will do for a client all year.

3. CAP and the ASA — you share responsibility for your clients' ads

The CAP Code is the rule book for non-broadcast ads, sales promotions and direct marketing, and it sets rules for advertisers, agencies and media owners to follow [6]. Rule 1.8 states that primary responsibility for observing the Code falls on marketers, and that agencies, publishers and other suppliers involved in preparing or publishing marketing communications also accept obligations under it [7]. If you write the copy for a claim you cannot substantiate, the fact that the client approved it does not remove you from the ruling.

If you add influencer or affiliate work, the disclosure rules are where agencies get caught. The ASA treats affiliate marketing as advertising falling under the advertising rules, requires that the commercial nature be clear before the user clicks through — an "Ad" identifier in a title, or on the image itself where the image appears first — and where only some links are affiliated, each affiliate link and its related content must be identified [8]. Crucially, allowing an affiliate control over the content of ads does not absolve the business of responsibility, and the ASA has held both the business and the affiliate marketer responsible for meeting the rules [8]. The ASA's influencer guidance is written for influencers, brands and agencies together, and treats all three as accountable for ad labelling and disclosure of commercial relationships [9]. Build the disclosure requirement into the influencer brief and the contract, and check the posts. Do not delegate it to the creator's judgement.

None of the above is a substitute for advice on your actual contracts. The VAT treatment of client ad spend in particular depends on the precise arrangement and must be confirmed with an accountant, and your processor terms should be reviewed by a solicitor before the first client signs.

Risks & Common Pitfalls

• Client concentration — the risk that actually ends agencies like this one. £3,500 of £5,300 is 66% in one place. If that client leaves in month four you are at £1,800 a month, which is below a living wage and below your break-even. Founding clients who signed on a relationship are not more loyal than clients who signed on a pitch; they are often less, because the relationship makes the conversation harder to have. Target three clients by month nine, not because you need the money but because you need the distribution.
• The ad-spend float. Covered in section 06 and section 12. Funding £45,000 a month against £18,000 on 45-day terms is the fastest route from profitable to insolvent available to you.
• The payment-terms gap, even without the media. A freelancer invoices you on 14 days while your client pays on 45. That is a 31-day hole of roughly £3,000 per freelance-delivered client-month, and two of them exceeds the £5,053 you have left. The fix is structural: bill retainers in advance, negotiate 30-day freelance terms, and never let the two sit on different sides of a month end.
• The capacity wall. Twelve delivery days a month, nine already spent. The third £3,500 client needs six days you do not have, and buying them from a freelancer at £500 a day leaves a 14% gross margin [1]. You will hit this wall about four months after it becomes obvious, so raise prices before you need to rather than after.
• You are the product and the single point of failure. Two weeks of illness in November is a client-losing event. The £5,500 of freelance cover in section 06 should be a named person you have already worked with, not a search you begin in a crisis.
• Correlated seasonality. Every client in an outdoor and cycling vertical peaks in the same months. A specialism concentrates your expertise and also your revenue curve. Plan cash for a flat February across the whole book, not one client at a time.
• Scope creep dressed as goodwill. "Could you just look at the website too" is how a £589 effective day rate becomes £400.
• Compliance drift. The processor agreement you never signed, the client list you never questioned, the influencer post you never checked. Each is cheap to do right at the start of a relationship and expensive to retrofit — and after February 2026 the PECR downside is materially larger than it used to be [2].
• Getting the VAT treatment of client media wrong. Not a risk you should be carrying personally. Put it to an accountant, in writing, before the first invoice [3].

Action Plan: 30 / 60 / 90 Days

The verdict is start it, so this plan assumes you do. The ordering is deliberate: the two things that can break the business are settled in the first thirty days, before you have anything at risk.

Days 1–30 — remove the two structural risks, then incorporate

• Ask both clients to hold their own platform billing. Get it in writing. This is the single highest-value action in the plan.
• Take the exact wording of both engagements to an accountant and settle the VAT treatment of any media you do touch, before you invoice [1].
• Incorporate digitally with Companies House, £100, and diarise the confirmation statement for twelve months out [2].
• Commission one master services agreement and one Article 28 processor agreement you can reuse for every client [3].
• Pay the ICO data protection fee — tier 1, £52, or £47 by direct debit [4].
• Business bank account, professional indemnity and cyber insurance, password manager with per-client vaults.
• Negotiate the larger client from 45-day to 30-day terms, or 14 days in advance. Trade fee for terms if you have to.

Days 31–60 — make the delivery repeatable

• Register for VAT voluntarily if your accountant agrees, and get digital record-keeping set up from the first return [5].
• Build the one-page monthly report once and use it for both clients.
• Document the onboarding checklist while you still remember what onboarding cost you.
• Write the sub-processor list and the two-page security summary. Procurement will ask.
• Identify and actually work with one freelancer for two days, so your cover is a person and not a plan. Back-to-back processor terms before they touch any data [3].
• Confirm how each client's email list was built, in writing, before your first send [6].
• Agree the 90-day metric and written case-study permission with both clients.

Days 61–90 — build the pipeline you do not have

• Launch the £1,200 fixed-fee audit as your front door, and publish two teardowns of brands you do not work with.
• Open five partner conversations — Shopify agencies, 3PLs, email platform partner managers. Referral, not reciprocity. Book one trade event in your vertical for the next six months.
• At the 90-day review, ask each client for one introduction. Put it on the agenda.
• Run the capacity number honestly: delivery days used versus twelve. If you are above ten, the next client is a price rise, not a new logo.
• At day 90, decide the year-one target. Getting from £63,600 to £90,000 needs one new client and one renewal increase, not a growth strategy.