Most studios with no marketing budget do not have a marketing problem. They have a publishing problem: twenty projects delivered, four of them shown, badly, as flat images with no story attached. This is not an article about what a marketing budget should be, or what an hour of your own time is worth; it assumes the number is zero and hands you the sequence, the scripts and the ninety day plan instead. Everything here is ordered by return per hour of your own labour, because that is the only currency you have, and the ordering matters more than the tactics. The first two sections are the whole plan if you only ever read that far.
What a zero budget plan actually is
Zero budget does not mean zero cost. It means you are paying in hours instead of cash, and hours are the scarcer resource when you are also delivering client work. So the discipline is not “do everything free”, it is “spend six hours a week on the two or three activities that return the most per hour, and refuse the rest without guilt”. Deciding what a studio should spend, or pricing your own hours honestly, is a separate calculation with its own arithmetic. This article starts one step after that, at the point where the answer is zero and the work still has to come from somewhere this quarter.
The compounding argument
Paid acquisition is a tap. It runs while you pay and stops the day the card declines. Everything in this playbook is an asset. A case study you publish in March is still being read in year three, still ranking, still being forwarded by a former client to someone who asked for a recommendation. The curve is flat for months and then steps, which is why most people quit around month two.
If you need a project inside thirty days, this is the wrong article: work your existing network directly, take a subcontract, or spend money. This playbook builds a machine that produces enquiries whether or not you feed it this month.
What you are trading away
Speed and predictability. Paid channels give you a dial. Organic gives you a ratchet. You cannot turn a case study up by 40 percent next Tuesday. What you get instead is a lower cost per enquiry over time, better-qualified enquiries (people who arrive having read three of your pages arrive pre-sold), and no dependency on a platform that can reprice you overnight. This article does not re-derive marketing budgets or acquisition costs; if you are trying to work out what to charge so the numbers hold up, the arithmetic in our piece on what a custom website costs is the place to start.
Start with the work you have already done
This is the highest return activity available to you and almost nobody does it. You have finished projects sitting in an archive folder. Each one is a fully paid-for asset that you have never shipped. Publishing it costs you a few hours and no money, and it works on every other channel at once: it is the thing search finds, the thing you send in outreach, the thing a partner forwards, the thing a referral checks before replying.
The arithmetic that makes this obvious
Assume a case study takes six hours end to end: an hour gathering assets, two hours writing, two hours on the before and after visuals, one hour building the page. Assume it is read by 400 people over twelve months across search, LinkedIn, referrals and outreach. Assume two percent of those readers send an enquiry, so eight enquiries. Assume one in four becomes a proposal and half of those close, so one project. At an average project value of 6,000 US dollars, that is 6,000 dollars from six hours of work, and it keeps running in year two. Every input there is an assumption you should replace with your own, but even if you halve the traffic and halve the conversion, it beats anything else on this list per hour.
Now run the same maths across five archived projects. Thirty hours, spread over five weeks at six hours a week. That is the whole first phase of your plan.
Getting permission without stalling
Permission is the usual excuse for not publishing, and it is mostly imaginary. Ask by email, be specific, and make saying yes cheap: “We would like to publish a write-up of the booking flow work. I will send you the draft before anything goes live and I will remove anything you are not comfortable with.” In our experience most clients say yes. Some say yes with conditions, which is fine.
If a client says no, or there is a real NDA, publish it anonymised. “A forty person logistics operator in Rotterdam” carries almost the same weight as the name, because the reader is evaluating the problem, not the logo. Change the brand colours, blur the company name in the screens, keep the interface and the numbers. Anonymised beats nothing, and studios who insist otherwise are usually avoiding the writing.
The ninety minute version
If six hours is not available this week, publish one screen. Pick a single interface you fixed, show the before and the after side by side at the same crop and scale, and write four hundred words about what was wrong and what you changed. That is a real artefact. Ten of those beat one perfect case study you never finish.
What a case study must contain to convert
A gallery of finished screens is decoration. A case study is an argument. The reader is trying to answer one question: would this studio solve my problem. Everything on the page either helps them answer it or is in the way.
The eight parts
Start with who the client is and what they sell, in two sentences, because the reader is scanning for someone like themselves. Then the problem in the client’s own words, with the commercial cost attached: not “the site felt outdated” but “the site was built in 2016, the booking form dropped anyone on a mid-range Android, and the sales team was rekeying enquiries from voicemail”.
Then the constraint, which is the part everyone leaves out and buyers care about most: budget, deadline, a legacy platform you could not replace, one part-time marketer as the only internal resource. Constraints are what make the work believable.
Then what you actually did, including the decisions and the options you rejected. Then the evidence: before and after, same crop, same device frame, same scale. Then the result, whatever you honestly have. If you have numbers, use them and say how they were measured. If you do not, say what changed and quote the client. Never invent a percentage; one fabricated uplift figure poisons the whole page for anyone who reads carefully, and the people who read carefully are the ones with budget.
Then the mechanics: stack, timeline, team size, who did what. Buyers use this to work out whether you can handle their thing. Finally, one next step, a single line and a single link, not a bank of buttons. Our write-ups for Pulse Clinic and US Construction follow this shape, and the sections that get the most attention are always the constraint and the before and after.
Why before and after beats a gallery
A gallery asks the buyer to judge taste. Most buyers are not qualified to judge taste and know it, so they fall back on price. A before and after asks them to judge change, which they are completely qualified to do, and it moves the conversation from “is this pretty” to “can you fix mine”.
There is a second reason. Almost every buyer already has a website, an app, or a brand. They are not commissioning a greenfield fantasy, they are commissioning a repair. A before and after is literal proof that you can work inside an existing business with existing constraints, which a beautiful concept shot can never demonstrate. If you show only greenfield work, you attract greenfield enquiries, which are rarer and usually worse funded.
Show the before honestly. Do not screenshot the old site on a broken viewport to make it look worse. Buyers who are about to hand you their own dated site will recognise the trick and infer how you will talk about them later.
Make your own site the proof, not a brochure
Your site is the only asset in this list you fully own. Everything else rents attention from a platform. So it should not read like a brochure that describes your services in adjectives; it should function as the evidence locker.
The three jobs of a studio homepage
One: say what you do and who you do it for, in plain language, above the fold, without a metaphor. Two: prove it with work, immediately, not four scrolls down. Three: give one obvious next step. That is the whole specification. The parallax hero with a rotating list of nouns is costing you enquiries.
Cut the pages that do nothing
The six-icon process section that says Discover, Define, Design, Develop, Deploy, Delight is dead weight on every studio site that has it. Either replace it with what actually happens in week one of your engagements, with dates and named deliverables, or delete it. The same goes for the services page that lists nineteen capabilities: a studio that does everything reads as a studio that has done nothing repeatedly.
Write the about page as evidence rather than personality, because for a small studio the question behind it is “will these people still exist in six months and can they run a project”, and the pattern that answers that is covered in the about page that builds trust. Fix the copy everywhere else too; most studio sites lose more to vague writing than to visual design, and the principles in our website copywriting guide apply to your own site before anyone else’s.
Say something about money
Put a starting range or a typical project band on the site. You will lose enquiries. Those are the enquiries that were going to waste two calls and then ask if you can do it for 800 dollars. Publishing a floor is the cheapest qualification mechanism you have, and it makes you the rare studio that treats the buyer’s time as real.
Search: win specific queries before head terms
A young domain with no links does not rank for “web design agency” and will not for years. Chasing head terms with a blog full of general advice is a waste of the only hours you have. What a young domain can win is the long, specific, low-competition query, and those queries are where buying intent actually lives.
Target the language buyers use, not the language you use
You say “legacy modernisation”. They search “our website looks old”. You say “conversion rate optimisation”. They search “why is nobody filling in my contact form”. Write down every question a client has asked you in a first call this year, verbatim, and treat that list as your content plan. It is better than any keyword tool you cannot afford.
Three families work reliably. Comparison queries: X versus Y for a specific situation, which buyers search the week before they commit to a platform. Cost and process queries: what something costs, how long it takes, what happens in what order. Problem and symptom queries: the thing that is broken, described the way a non-designer describes it. A page like how to choose a web design agency exists because that is a real thing buyers type days before they contact anyone.
The arithmetic of small queries
A single specific query might get forty searches a month. Rank third, take ten percent of them, and you get four visits a month, about fifty a year. At a two percent enquiry rate that is one enquiry a year. Per page, that sounds pointless. That is why people give up. Now build twenty-five of those pages over a year: twenty-five enquiries, at a cost of your writing hours, compounding as the domain ages and the better pages start ranking for terms you never targeted. Search only works in aggregate, and only if you commit to a year.
Write for the answer engines too
A growing share of buyers now ask an assistant rather than typing into a search box, and the pages that get quoted are the ones with a clear direct answer near the top, real specifics, and no throat-clearing. That is the same discipline as writing well for humans, with a few structural habits on top, which we cover in search in the AI era. Do not build a separate strategy for it. Write pages that answer the exact question and get to the point in the first paragraph.
Portfolio platforms are distribution, not portfolios
Dribbble, Behance, Awwwards, Layers, LinkedIn. Studios treat these as their portfolio. They are not a portfolio, they are a distribution channel, and the difference determines whether they return anything.
Own the destination
Never let the post be the terminus. Post an excerpt, an animation, one strong frame, and send the reader to the full case study on your own domain. The platform gets the attention, you get the visit, the analytics, the follow-on reading and the contact form. If your entire body of work lives inside someone else’s product, you are building a business on rented ground and you cannot see who is looking.
Be honest about the audience
Most people on design platforms are designers. They are not your buyers. What they are is a source of subcontracting work, referrals from studios who are at capacity, and hiring leads, which are all worth having but should not be confused with client acquisition. Agency overflow is one of the fastest paths to paid work for a small studio, and the studios most likely to hand it to you are the ones one size up from you, not the ones five sizes up.
LinkedIn is the exception, and only if you fix the input. If your connections are all designers, your posts reach designers. Spend twenty minutes a week connecting with operations managers, founders, and marketing leads in the two industries you want to work in, and the same post starts landing in front of people who sign cheques. Post the before and after image with three sentences of context and a link. That is the whole format.
Ask for referrals properly
Referrals convert better than anything else on this list for most studios, and you can verify that in an afternoon by listing where your last ten projects came from. They are also the channel studios handle worst, because the ask is vague and the timing is wrong. “If you know anyone who needs a website, send them my way” is not an ask. It is a wish. It puts the work of remembering and matching onto a busy person, so nothing happens.
The timing that works
There are three moments. The day something ships and visibly works. The moment a client says something nice unprompted, which is the strongest signal you will ever get and should be acted on within the hour. And about thirty days after launch, when the first real numbers are in and you are having the review conversation anyway, which is exactly the check-in described in our piece on the first thirty days after launch.
The moment that does not work is when you send the final invoice. Attaching a favour to a payment request makes both feel transactional.
The ask that works
Name the target and write the message for them. “You mentioned your friend at the other clinic was rebuilding their booking system. Would you be open to introducing us? I have written the email below so you can forward it without thinking about it.” Then three sentences they can forward verbatim, with the case study link in it.
Run the numbers on your own list. Twelve past clients, eight of whom you can reasonably ask, three of whom actually make an introduction, one of which becomes a project. That is a couple of hours of email for a project, and it is repeatable roughly every six months without becoming a nuisance.
Ask for the testimonial as a separate conversation, not bundled with the referral, and ask for something specific: “Could you say a line about what the enquiry volume looked like before and after?” Then place those quotes where they do work rather than in a carousel nobody reads, using the placement logic in our piece on social proof on websites.
Stay in view after the project ends
Most referrals die because you disappeared. A two-line email every quarter to every past client, with no ask in it, is enough: here is something we shipped, here is one thing we learned that applies to you. Studios that keep a light retainer or a maintenance arrangement in place get referrals continuously, because they are still in the room when someone at that client’s next company asks who built their site.
Partnerships with adjacent suppliers
The people who already sell to your buyers, and do not compete with you, are the cheapest distribution you will ever find. They have trust you have not earned yet and they encounter your exact buyer at the exact moment the need appears.
Who to approach
Developers who do not design and hate being asked to. Copywriters who keep getting handed terrible layouts. Commercial photographers who shoot for the same businesses. Small PR firms. Accountants and fractional finance people who serve owner-run companies. Webflow and Shopify implementers. Brand strategists with no execution arm. Other studios with a full pipeline. Aim for six real relationships, not sixty logos.
Make the first move worth something
Nobody wants a coffee to “explore synergies”. Turn up with something. Do a free thirty minute review of one of their client sites and send them the notes so they can pass it on and look good. Offer to be the design half of a proposal they are writing this month. Send them a lead you cannot take. Give first, three times, before asking anything, and be willing to accept that one or two of the six will never reciprocate.
The artefact that makes partners useful is a short, specific audit they can hand to a client, which is the same document described in our piece on auditing a site before a redesign. Give them the template so they can spot the problem, then bring you in.
Skip the formal partner programme
Do not build a partner portal, a tier structure, or a commission scheme when you are three people. A ten percent referral fee turns a peer into a salesperson, taints the recommendation in the client’s eyes if it ever comes out, and creates admin you will resent. A genuine relationship with reciprocal work, a named contact, and fast responses beats it every time. Paying for cold introductions from people with no standing in your market is money and hours thrown away.
Direct outreach that is welcome
Cold outreach has a bad name because most of it is automated garbage. Done at low volume with real research, it is one of the few channels that works from a standing start, before you have search rankings or a referral base.
The bad version
“Hi there, I hope this email finds you well. We are a leading design agency delivering world-class digital experiences for forward-thinking brands. I noticed your website could use a refresh. Do you have 15 minutes on Thursday for a quick call? Here is my calendar link.”
Everything is wrong. No evidence of having looked at the business. “Could use a refresh” is an insult with no specifics. The ask is a meeting, which costs the recipient thirty minutes of their life for your benefit. The calendar link presumes a relationship. It could have been sent to nine hundred people, and the recipient knows it.
The good version
Subject: your quote form on mobile
“Adam, I went through your quote request form on an Android phone yesterday, on the assumption that most of your traffic is mobile. The postcode field rejects lowercase input and the error message appears above the fold you have already scrolled past, so it looks like nothing happened when you press submit. I recorded a two minute screen capture of it. Happy to send it over, no strings; you can hand it to whoever maintains the site. We rebuilt a similar quote flow for a construction firm last year and the write-up is here if it is useful.”
Ninety words. It names one concrete defect that costs him money. It proves the work was done before the ask. The ask is “may I send you a video”, which costs him five seconds and gives him something valuable. There is no calendar link and no attachment. The case study is offered, not insisted on. Assume one in five replies; that is the number to test against your own list, not a promise.
Two more rules. Only send it when the observation is real; a manufactured problem gets you correctly filed as spam. And find a trigger: a funding announcement, a new product line, a new marketing hire, a second location, an event they are exhibiting at. The trigger is the answer to “why are you emailing me now”, and without it the message floats. If the business is trading well and the site is working, the risk of changing it is the real objection, and having a clear position on redesigning without losing conversions is what turns that objection into a conversation.
Volume and cadence
Ten to fifteen researched emails a week, roughly fifteen minutes of research each. Not two hundred. One follow-up after eight days, one line long, then stop forever. Fifty emails a month, a twenty percent reply rate, ten replies, three real conversations, one project per quarter. If those numbers hold for you, the channel is paying for itself at about four hours a week. If you send eight hundred templated emails instead, you get a burned domain and nothing.
Community participation
Being visibly useful where your buyers already gather is slow, and it is the only channel on this list that costs nothing but attention. The mistake is going where designers are. Your buyers are in industry Slack groups, founder communities, local business associations, trade forums, and niche subreddits about running a clinic or a construction firm or a SaaS product.
Answer the question and stop
Answer completely. Give the actual advice, including the part that means they do not need to hire anyone. Do not link to yourself. Your profile is enough; people who want to know who you are will click it, and the ones who do click arrive with more goodwill than any advertisement can buy. One thorough answer a day, capped at twenty minutes, is the right dose.
Show up in person where it is cheap
A local business meetup, a founders’ breakfast, a small industry conference where you are not speaking. Offer to do a free teardown night: five businesses bring their site, you spend fifteen minutes on each in front of the room. It costs an evening, it demonstrates exactly what you are like to work with, and the people in the room remember it for months.
Turn one project into several artefacts
Every finished project should produce more than one thing. The writing is already done once you have the case study; the rest is packaging.
The ladder from a single project
From one redesign you get: the case study on your own site; a before and after image set for social; a short screen recording of the new flow; a written insight about the one hard problem you solved, which becomes a search page; a LinkedIn post with the strongest frame; a platform entry linking back; a checklist or template you used internally, cleaned up and published; a short talk for a meetup; and a paragraph in your quarterly email to past clients. Nine artefacts, maybe five extra hours on top of the case study.
The rule that makes this work: every artefact points at the same destination on your domain. If the LinkedIn post, the platform entry and the talk all send people to the same page, that page accumulates authority, links and evidence instead of scattering attention across nine dead ends. Nine artefacts pointing at nine different pages build nothing; the authority, the links and the return visits all land in separate places and none of them accumulates.
A weekly schedule that fits around client work
Six hours. If the plan needs more than that, it will collapse the first time a project goes sideways, and a plan that collapses is worth nothing.
Monday, ninety minutes: publishing. One case study or one search page, moved forward. This slot is protected; it is the highest return hour you have.
Tuesday, forty-five minutes: outreach research. Build the list, find the triggers, write the observations. Do not send anything.
Wednesday, forty-five minutes: send the outreach, clear replies, book calls.
Thursday, sixty minutes: relationships. Two conversations, whether that is a partner, a past client, a referral thank-you, or a quarterly check-in email to five people.
Friday, sixty minutes: repurposing and posting. Cut the artefacts from whatever shipped, post them, link back.
Any time, thirty minutes: community answers, split across the week.
That is five hours and thirty minutes, plus a two hour review at the end of every month where you look at what produced conversations and cut something.
What to cut when the week collapses
Cut the platform posting first. Cut community second. Cut outreach third. Never cut the Monday publishing slot, because it is the only one that keeps returning after you stop.
The first ninety days from zero
Days one to thirty: build the evidence. Inventory every project you have ever delivered. Email every client you need permission from on day one, before you start writing, because the replies take a fortnight. Publish three case studies. Rebuild your own site around the work rather than the services. Install analytics and decide now what a success looks like, using the short list of numbers in our piece on what to actually measure. Do no outreach this month. You have nothing to point at yet.
Days thirty-one to sixty: distribution. Publish two more case studies and four search pages drawn from questions real clients have asked you. Make the referral ask to eight past clients, individually, with the forwardable paragraph written for them. Open five partner conversations and give something in each one before asking for anything. Start posting excerpts on one platform and one social channel, both linking back.
Days sixty-one to ninety: outreach and compounding. Now start the ten to fifteen researched emails a week, because now every message can point at real proof. Keep publishing weekly. Run the first repurposing cycle properly on your best project. At day ninety, review honestly: count the conversations, not the impressions, and note where each one came from. Kill whichever channel produced nothing and put those hours into whichever produced something.
Expect the first project attributable to this work somewhere between month three and month six. If it lands sooner it will almost certainly come from the archive you published in month one, or from a past client you re-contacted, not from anything you started in month three.
What to measure, what to skip, and what this will not do
Two numbers matter early: qualified conversations per month, and the origin of each one. Ask every enquiry how they found you, in the first call rather than as a required form field, and write the answer down. Attribution will be messy because people read three things over five months and then mention the last one. Treat the answers as directional, not as truth.
Skip these, actively: automated LinkedIn connection sequences, follower-count games, paid directory listings, awards you have to pay to enter, generic blog volume with no query behind it, hiring an SEO consultant while your domain is six months old, and rebranding your own studio instead of publishing your work. The last one is the most seductive and the most expensive, because it feels like progress and produces no evidence.
The honest summary: this is slower than paid acquisition, it demands consistency for months before it pays, and the first sixty days feel like shouting into a void. What it gives you in return is an asset base that keeps working when you are heads-down on delivery, enquiries from people who have already read enough to trust you, and independence from any channel that can change its pricing on a Tuesday.
If you are working through this and want a second opinion on which of your archived projects to publish first, or on whether your own site is doing the proof work it should, write to us at hello@beconfidency.agency or send the list through the contact page, and we will tell you what we would do in your position. We run the same playbook on ourselves, so the answer will be specific rather than theoretical, and we are happy to be blunt about the parts we got wrong.
When the constraint stops being your archive and starts being the site the archive sits on, the shape of the fix is in the work we publish.
When the archive is ready and the site under it is not, that is exactly what our SEO and performance service is for.