How to Find Clients: Six Channels and What Works Fastest
Finding clients means picking a channel that reliably puts a person with a budget and a problem in front of you, then working it every week instead of in bursts. There are six: paid ads, cold outreach, referrals, content, marketplaces and existing demand — people writing in public right now that they need what you sell. The last one produces the first client fastest, because the request is already written for you.
- how to find clients
- how to attract customers
- customer acquisition channels
- where to find clients
- how to find your first clients
- getting clients for a small business
The short answer: where clients actually come from
Clients come from six places: you pay for attention (ads), you write first (cold outreach), someone vouches for you (referrals), people find you through your own material (content), you stand on somebody else's shelf (marketplaces), or you go to someone who has already said out loud what they need (existing demand).
You pick between them on three questions, not on abstract effectiveness: what the channel costs, how long until the first client, and what you can do in it today with no budget. If you need a client this week, the choice narrows to two — pay for impressions, or go where the demand is already written down in words.
Ads — fast, but you pay per impression and it stops when the budget stops.
Cold outreach — cheap, but reply rates keep falling and reputation burns fast.
Referrals — the warmest channel, but you cannot switch it on, only grow it.
Content — it works for years, and gives you almost nothing for six months.
Marketplaces — briefs arrive at once, for a commission, next to cheaper rivals.
Existing demand — someone already wrote what they need; you find it and reply.
Customer acquisition channels: cost, timeline, first step
There is deliberately no average-cost-per-lead column: the spread between niches runs to tens of times, and any round number here would be invented. Instead you get what can genuinely be compared — cost, timeline, and the step available today. Read the timeline column first: if you need revenue this month, anything that pays off in six months is background work, not a choice.
Content and marketplaces are the two extremes. Content returns almost nothing for six months, then one article keeps working for years at no extra cost. Marketplaces send briefs immediately, but the client belongs to the platform: a commission, cheaper neighbours, and no say in whether you are shown tomorrow.
| Channel | What the money goes on | Time to first client | What you can do today |
|---|---|---|---|
| Paid ads | Budget for impressions and clicks, plus creative work | Days — but only while you keep paying | Write one offer and run a small test budget against it |
| Cold outreach and calls | Time on lists and copy, plus reputation risk | Weeks, and only at real volume | Build a list of twenty by hand and write each one differently |
| Referrals and word of mouth | Almost nothing beyond good work and asking at the right moment | Months — it grows with the number of happy clients | Message three past clients and ask them directly for an introduction |
| Content and social | Consistent author time, month after month | Six months and beyond, then it compounds on its own | Write down the five questions clients ask you most often |
| Marketplaces and freelance platforms | Platform commission and constant price pressure from neighbours | Days, but the client belongs to the platform | Finish your profile and reply to five fresh briefs |
| Existing demand in open discussions | Your time on searching and on direct messages | Days — the request is already written by the buyer | Find today’s “looking for…” messages in your category and answer three |
Ads: when they pay off and when they are just expensive
Ads are the only channel with an on switch. That is both the strength and the trap: while the money flows, enquiries flow; when it stops, so do they, and you are exactly where you started.
They pay off in two situations: you have repeat purchases, so you can pay more for a client than the first sale returns; or your offer is narrow and you know where that audience sits. Otherwise ads pay for other people's curiosity. And do not switch them on before the offer is tested — until you know which sentence makes a person reply, you are paying to find that out.
Count money, not enquiries: what a client costs and what they are worth over the whole relationship.
Test the offer in conversation before you pay for impressions.
Do not mix several audiences into one ad, or you will not know what worked.
Budget a month for testing — the first combination is almost never the one that works.
Cold outreach: why reply rates keep falling
Cold outreach used to work on volume: a thousand emails, ten replies. The same volume returns much less today. Everyone sends those thousands automatically, people have built up immunity, and filters catch identical messages in batches.
There is also a cost that never shows up in the report: mass sending burns the list, and someone you wrote to off-topic will not read you a year later either. In messengers the price is higher — identical messages to strangers get an account restricted, so you lose a working tool, not just a campaign.
Twenty different messages will beat a thousand identical ones.
The reason you are writing has to be visible in the first line, or it reads as spam.
Identical messages to strangers in messengers are the shortest route to account limits.
Do not send from your main working account: getting it back costs more than the campaign.
Referrals: how to get them on purpose
A referral is the cheapest and warmest client you will get, because trust arrives attached to the contact. But most people wait passively — do good work and they will talk about us. They will. Rarely, and not in the month you need money.
Referrals become a channel the moment asking becomes a habit. Project finished, so you ask — not "keep me in mind", but specifically: "is there anyone you know with the same job on their plate right now?" The second source is people who serve your clients without competing with you: designer and developer, accountant and lawyer. Two or three such partners make the flow predictable; the same logic is laid out for agencies.
Ask while the client is happy — right after the result, not six months later.
Keep the question narrow: "who do you know that needs this right now?"
Find two or three adjacent suppliers and agree to trade referrals.
Record who referred whom, or you will not know who to thank.
See also: How agencies run this
Existing demand: people who already wrote what they need
Some demand does not need creating, because it is already stated. Every day in open professional chats, city communities and industry discussions, people write it plainly: "looking for a contractor", "can anyone recommend a tool", "who has done this, share a contact". That is a brief, written by your future client in their own words.
That person costs less than any ad click, because the step where you explain why the thing is needed is already done. The only hard part is volume: open chats produce thousands of messages a day and the useful ones sink between memes and off-topic.
That is the part Leadgram takes over: it reads public messages by meaning rather than exact word match, and returns people with a 0-100 buying-intent score, a short reason, and the message itself. For Telegram specifically, see finding clients on Telegram.
The buyer writes the brief, so you are not guessing at the pain or the audience.
The timing is right: they are looking now, not "sometime later".
You see context — the full message, the group it came from, and what was said around it.
See also: Finding clients on TelegramHow the 0-100 score works
The cheapest demand is the demand that already exists and is already written down. You do not have to create it with ads — you have to notice it in time.
How to search open chats for stated demand
Telegram is where these messages sit most densely: industry chats, community chats, local classifieds, channels with open comments. All of it is public — what any person who opens the chat can see. Leadgram works only with those messages and never exports member lists. Reading what someone said out loud and hoovering up everyone in a group are not the same thing.
Start from the query rather than the chats. Describe who you are looking for in ordinary words, the way the person themselves would say it: "looking for someone to build a landing page", "need an accountant for a small company". Then read the top of the list rather than the size of it.
Collect three to five phrasings the way a client would write them, not the way you name your service.
Run the search — no login or signup required, and the first results appear immediately.
Sort by score and read the source messages at the top: the score narrows it, you decide.
Drop anything asked six months ago: freshness beats volume here.
Save a query that worked, and add your own chats as sources.
See also: Search channels and groupsSaved searchesYour own group sources
How to write first and actually get a reply
Finding the person is easier than getting an answer. One rule carries most of the weight: the first message has to be obviously written for that person and about their job. A single line referring to what they wrote is worth more than a paragraph about your company.
The structure that almost always works: refer to what they wrote, say in one sentence how you can help, ask one easy question. What reliably ends the conversation is a bare "Hi!", a wall of text, a link instead of an answer, and selling before the person has said a word.
First line about their message, not about you.
One sentence on how you help — concrete, no generalities.
One question at the end that can be answered in a word.
No price list, no deck, no website link in message one.
No reply? One short follow-up a few days later, and that is it.
See also: Conversations in one inbox
How not to lose the people you found: stages, reminders, numbers
Half of all clients are lost after the search, not during it. Someone replies, asks you to send an example, you get pulled elsewhere, and two weeks later they are working with somebody else. A person you found is still a lead, not a client, and one rule fixes that: every lead has a stage, an owner and a date for the next step.
Leadgram has a board for exactly this: stages Lead → Contacted → Qualified → Won → Lost, which you can rename, reorder and extend with your own, plus deal value, notes, a change history and reminders. Leads land on it straight from search, carrying their score and source message.
In the first month, four numbers are enough: how many you wrote to, how many replied, how many reached a conversation about money, how many bought. That is a sales funnel at its simplest — enough to see where things break, almost always at the first message.
Everyone gets a stage: "thinking about it" is a stage, not a lost contact.
The date of the next step matters more than a tidily filled-in card.
Once a week, work through everything that has not moved in ten days.
Need numbers outside the board? Export the list to CSV.
See also: Pipeline board and remindersWhat a lead isSales funnelCSV export
Frequently asked questions
Where do I find clients if I have no ad budget?
Where demand is already stated in words: open professional and local chats, and comments under topical posts. Plus two free channels most people never get round to — past clients you can ask directly for an introduction, and adjacent suppliers who serve the same audience. A Leadgram search runs without login or signup, and the free beta plan gives 50 ranked leads a month with no card.
How do I find clients on Telegram?
Not by exporting group members — that gives you a list of people with no sign that any of them need anything. The working approach is the reverse: read what people write in open chats and look for messages that carry buying intent, like "looking for" or "can anyone recommend". Leadgram does that with semantic search. There is a full page on finding clients on Telegram.
How much does one client cost from each channel?
There is no universal number: the spread between niches runs to tens of times. Work it out for yourself — divide what you spent on a channel this month by the deals that came from it. Paid ads usually cost the most, referrals the least, and existing demand costs only your time on searching and messaging.
How do I write a first message that does not look like spam?
Refer to what the person wrote themselves, explain in one sentence how you can help, and ask one short question. No price list, no deck, no link. The tell for spam is not length — it is the feeling that the same text went to a hundred other people.
How do I tell whether someone is really ready to buy?
From their words, not their profile. Readiness shows in the phrasing: a specific job named, a deadline, a mention of budget or of a supplier who did not work out, and the question asked in the present tense. Leadgram scores exactly that, 0-100 with a short reason — but you decide, because the original message is always shown next to it.
How many channels should I run at once?
Two. One fast channel that brings money now — existing demand or ads. And one slow channel that accumulates — referrals or content. Three at once nearly always means none of them is done consistently.
Find your next leads in Telegram
Run a search, review scored matches with the reason they fit and the source group, and export a clean list — all from public signal, no Telegram login.