Lead generation for agencies: how agencies actually get clients
The nine channels that bring agencies work, the economics of each, and how to build a pipeline without hiring a sales team.
Published by AI Knows Us · Updated 12 October 2026
Most agencies get most of their work from referrals, past clients and partnerships, which is why the channel that deserves your first week of effort is the one you already have. The rest of the pipeline comes from a deliberate second channel chosen to match your positioning: outbound for a narrow, nameable target list, content and search for a category people search for, partnerships for white label capacity, events for high value relationships. The decision is not which channel is best in general; it is which one your economics support and your team will do every week without being chased.
This page covers the three numbers that make the decision, the nine channels with the honest economics of each, and a weekly rhythm that works without a sales team.
Three numbers decide which channels you can afford
- Average contract value, counted over the whole engagement rather than the first month. A channel that costs too much for a three month project is cheap for a two year retainer.
- Gross margin, after the delivery cost of the work. An agency at a thin margin cannot fund a long payback channel, whatever the revenue looks like.
- Average retention, in months. This is the number most agencies have never calculated, and it changes every acquisition decision. Improving retention by a few months is usually cheaper than opening a new channel.
Then one piece of arithmetic, done with your own figures rather than anybody's benchmark. To find how many real conversations you need, divide your new revenue target by your average contract value, then divide that by your close rate on qualified conversations. Work back through your own ratios from conversation to proposal to win, and you have a monthly activity number that is actually yours.
The nine channels, honestly
| Channel | How it works | Time to first client | Where it fails |
|---|---|---|---|
| Referrals and past clients | Asking, systematically, at the moments when a client is pleased. A monthly check in list, not a hope | Weeks | Nobody owns the asking, so it happens when work is short and not before |
| Partnerships and white label | Other agencies and consultancies subcontract the work you are best at | One to three months | Margin is lower and the relationship can end with one hire on their side |
| Outbound email and LinkedIn | A named list, a specific reason to contact each one, a short message about them | One to two months | Volume with no research. Also the fastest way to damage a brand you spent years building |
| Content and search | Pages that answer what your buyers search, compounding over quarters | Three to nine months | Publishing generic advice instead of the specific thing you are asked in sales calls |
| Paid search | Buying the high intent phrases, such as your service plus a city | Days | Clicks arriving at a home page instead of a page about that exact service |
| Events and speaking | Industry events, trade bodies, local business groups, your own small roundtables | One to six months | Attending without a follow up plan, so the cost is spent on the room and not the relationships |
| Directories and marketplaces | Listings and reviews on platforms buyers shortlist from | Weeks to months | Paying for placement without a profile anybody would pick, and no reviews |
| A productised offer | A fixed scope, fixed price piece of work that starts the relationship: an audit, a sprint, a migration | Weeks | Pricing it so low that it is a loss leader with no path into a retainer |
| Community, podcast or newsletter | A regular audience of the people who buy what you sell | Six months and up | Stopping at month four, which is where nearly all of them stop |
What to do first, second and third
First, make referrals a process. List every past client, every current contact who has moved jobs, and every supplier who serves the same buyers. Contact each one with something useful rather than a request. Put the list on a monthly cycle with a named owner. This is the cheapest pipeline any agency has and the most commonly neglected.
Second, pick exactly one acquisition channel matched to how your buyers behave. If you can name the two hundred companies you want, outbound and events fit. If your service is something people search for, content and paid search fit. If your capability is specialist and other agencies lack it, partnerships fit. Commit for two quarters, because a channel abandoned at week six teaches you nothing.
Third, add a productised entry offer. A fixed scope paid audit or sprint converts far better than a proposal for a retainer, because the buyer is deciding on a small commitment. It also lets you see the account before pricing the long engagement, which protects your margin.
Positioning makes every channel cheaper
The single biggest lever on agency lead generation is not a channel, it is specificity. "We do digital marketing" competes with everybody and has to buy attention. "We do paid search for subscription box businesses" gets introduced, gets recommended, gets found and can charge more. Narrow by sector, by platform, by company stage or by a single outcome. You can always take the work outside the niche; what you cannot do is be recommended for something you have not named.
A practical test: ask three past clients to describe what you do in one sentence. If the three sentences do not match, prospects are guessing too.
A weekly rhythm without a sales team
- Monday, one hour: five referral or past client contacts, each with something specific to say.
- Tuesday, two hours: the chosen channel. Ten researched outbound messages, or one page written, or one event follow up sequence.
- Thursday, one hour: move every open opportunity forward by one step, and close the ones that are not real.
- Friday, fifteen minutes: update three numbers. Conversations this week, proposals out, revenue won. Three numbers on one sheet beats an unused pipeline tool.
- Monthly: review which channel produced the conversations, and cut anything that produced none for two months running.
This page covers how an agency fills its own pipeline. For how your clients should evaluate you from the other side of the table, see how to choose an SEO agency, which is a useful mirror to hold up to your own proposal. For building the content channel itself, see building a content marketing strategy.
Where people get stuck on lead generation for agencies
Should we hire a salesperson?
Hire when there is already a channel producing more conversations than the founders can handle, and when somebody can describe the sales process on one page. Hiring a salesperson to invent the channel rarely works, because the founder is still the one who can answer the hard questions in the room.
Does cold outbound still work for agencies?
It works at low volume with real research: a specific observation about that company, a relevant piece of proof, one clear ask. It stops working at the point it becomes a sequence sent to a purchased list. Check the marketing and privacy rules for every market you contact, and take local advice before you start a programme.
Should we buy leads from a lead generation provider?
Treat it as a test with a defined budget and a defined measure, and ask two questions before signing: how the leads are sourced, and how many other agencies receive the same lead. Where the answer is several, you are buying a race on price. Your own referral and content channels are worth more because they are exclusive.
How long before a content channel pays?
Plan on two to three quarters before it produces enquiries consistently, and fund it from a channel with a faster return in the meantime. The compounding is the point: the pages you wrote last year keep working while you write this year's.
Should an agency do its own SEO?
Yes, and treat it as the showcase. A prospect who finds you through the exact phrase you would be hired to win has already seen the proof. Run it as a real project with an owner and a deadline, because agency websites are the first thing to slip when client work is busy.
One question your prospects are already asking elsewhere
Of the nine channels, none lets you see what a prospect read before replying to your email. Increasingly, part of it was an assistant's answer to: which agency should I hire for this?
That answer behaves like a referral. It names a few agencies with reasons, and a prospect arrives already leaning. Ask it for your niche and city, without your name, and treat the result as a free competitor report. AI Knows Us runs the same question on a schedule so you see whether your positioning work changes who is named.