How to get clients: nine channels, ranked by what they actually cost

Every channel works for somebody, which is why channel advice on the internet is a shouting match between people who are all technically right. Cold email works — for someone with a list of 40,000 and a deliverability engineer. Referrals work — for someone twelve years into a career with a network to match. The only question that survives contact with your actual business is what a signed client costs you, with your own hours priced at what you could otherwise have billed. Most people never run that number, which is why most people are busy and broke at the same time.

1. Referrals

Highest close rate, lowest cost, worst scalability. The failure mode is not asking. Asking at the moment a client is visibly happy roughly doubles referral rates in most businesses.

2. Communities where your buyers ask

Local groups, industry groups, forums. Cost is your attention, and the leads arrive pre-warmed because someone asked publicly. Capped by how many conversations you can genuinely take part in.

3. Partnerships

Adjacent businesses serving the same client. A photographer and a wedding planner, a bookkeeper and a lawyer. Slow to establish, then durable and nearly free.

4. SEO and content

Expensive in time up front, close to free per lead once it works, and months before it does. Best where your buyers search for a solution rather than ask a neighbour.

5. Cold outreach

Scales with volume and infrastructure. Reply rates are low, deliverability is an ongoing engineering problem, and anti-spam rules keep tightening. Works far better for large addressable markets than for narrow ones.

6–9. Ads, marketplaces, events, directories

Paid ads buy speed and stop the moment you stop paying. Marketplaces trade margin for volume and train you to compete on price. Events convert extremely well per conversation and cost a day each. Directories are cheap, passive and low-intent — fine as a supplement, thin as a strategy.

How to actually choose

Pick two: one that produces clients this month and one that compounds. Running six channels badly is the most common reason a service business is busy, visible and still short of work.

The nine channels, compared honestly

Pick exactly two

One that produces clients this month and one that compounds. The most common failure is not choosing the wrong channel — it is running six of them at a quarter of the required effort, so none reaches the point where it works.

The same nine channels, by trade

Which two you pick changes completely with what you sell and how it gets chosen. A roofer and a copywriter are not running the same playbook, even though the channel list is identical. These break the choice down per trade, with the phrases buyers actually use and the groups where they ask.

Pick the one that matches what you do

• Trades and home services: contractors · roofers · HVAC · plumbers · handymen · landscaping · lawn care · pressure washing · cleaning
• Creative and freelance: photographers · designers · web designers · developers · copywriters · writers · social media managers · wedding planners
• Professional services and agencies: consultants · bookkeepers · financial advisors · law firms · marketing agencies · SEO agencies · recruiters · staffing agencies · virtual assistants · notaries · real estate agents · travel agents
• Health, coaching and care: therapists · life coaches · personal trainers · dog trainers · home care · daycares

Why channel advice contradicts itself so often

Most of it is written by someone generalising from the single channel that worked for them, in a market with characteristics they never mention. The two characteristics that decide almost everything are how often the purchase happens and how it gets chosen. A service bought once every few years and chosen by asking a neighbour — a roof, a wedding photographer, a divorce lawyer — lives or dies on referral and local visibility, and paid search mostly captures the small share who did not ask anyone. A service bought repeatedly and chosen by comparison — a contractor, an agency retainer — rewards being findable and having something to compare. Advice that works brilliantly in the second market fails completely in the first.

Referrals: the mechanics people skip

The reason referral programmes underperform is almost never the incentive. It is that the referrer does not know who to refer. “Let me know if you hear of anyone” asks them to hold an open-ended search in their head for months. “Do you know anyone else running a Shopify store doing over a million a year who is unhappy with their email?” gives them a query they can run against their own memory in three seconds. The specificity is the whole mechanism, and it costs nothing.

Communities: the capacity ceiling nobody plans for

Community-sourced work has a hard ceiling that is easy to miss because it arrives gradually. You can only reply to what you can read, and reply quality falls off sharply once you are skimming. In practice most solo operators max out somewhere around fifteen to twenty genuine replies a week before the answers get generic and the returns collapse. That ceiling is not a reason to avoid the channel — it is a reason to be in fewer, better-matched groups rather than in thirty.

Partnerships: why they fail and how to structure them

Most partnerships die because they are reciprocal by assumption. Two businesses agree to send each other work, one sends more, resentment follows, both stop. The version that survives is asymmetric on purpose: you refer them because it is genuinely the best outcome for your client, and you do it whether or not it comes back. That sounds naive and is the opposite — the businesses that get referred most are the ones with a reputation for referring well, and reputation is the asset that compounds here.

Cold outreach: what has changed

The mechanics have got harder in a way that is easy to underestimate. Bulk sender requirements introduced by the major providers now demand authenticated sending, low complaint rates and easy unsubscribe for anyone sending at volume. The practical effect is that cold email at scale is now an infrastructure discipline: separate domains, warmed inboxes, complaint monitoring. That is a real cost, and it should be counted before comparing the channel to anything else. It still works — but it works for teams who treat deliverability as an ongoing job, not for someone sending 200 messages from their main domain.

How to actually run the comparison

Pick one number: cost per signed client, with your own hours priced at what you could otherwise bill. Track it per channel for a quarter. Almost every argument about channels dissolves once both sides are measured the same way — and the results are frequently the reverse of the intuition, because “free” channels consume the resource you never invoice for.