How do I find where my target audience spends time?

Work backwards from the complaint rather than forwards from the demographic. Type the sentence a frustrated customer would type into a search engine and read what ranks: subreddits, trade forums, Facebook groups, review pages, YouTube comment sections. Then ask five existing customers where they looked before they found you. Those two lists overlap more than most owners expect, and the overlap is where to start.

Should my small business be on every social platform?

No. One platform posted on consistently beats six posted on occasionally, because every platform rewards frequency and punishes absence. Spreading a limited amount of attention across six accounts produces six dead accounts, each of which makes the business look abandoned to anyone who checks. Pick one, run it properly for ninety days, and add a second only when the first runs without your daily involvement.

What is the difference between a search channel and an interruption channel?

On a search channel the person arrives already looking for what you sell, so intent is high, competition is priced in, and volume is capped by how many people are searching. On an interruption channel you appear in a feed and have to create the intent yourself, so volume is nearly unlimited and quality is lower. Search converts better. Interruption tests faster and cheaper, which makes it the better place to learn what language works.

How do I know if a channel is working before I have spent months on it?

Decide the threshold in advance and run a fixed two-week test. Look for people who were not already connected to you asking a question about the work — not likes, not follows, not views. Three unprompted enquiries from strangers in two weeks is a real signal at small scale. Zero, from a channel where you posted honestly and specifically ten times, is also real information.

Where do local businesses find customers online?

Local businesses have a narrower and easier answer: map listings and local search first, neighbourhood groups second, and referral partners third. The population that can ever buy from you is bounded by driving distance, which makes broad platforms inefficient and makes reviews and local search unusually powerful. A well-maintained map listing with recent reviews often outperforms every social account combined.

What if my customers are not really online in a findable way?

Some are not. Buyers of industrial equipment, elderly homeowners, and certain trade purchasers gather at suppliers, trade counters, associations, and job sites rather than in public forums. In that case the online channel to use is search, because they will still look you up before calling, and the acquisition happens through referral partners and physical presence. That is a legitimate answer, not a failure to find them.

How long should I give a channel before deciding it does not work?

Ninety days of genuine consistency for an organic channel, and two to three weeks for a paid one. The difference is that paid buys attention immediately while organic has to accumulate it. What ruins most organic tests is not the ninety days but the seven-post start followed by a three-week silence, which measures your consistency rather than the channel.

Where are my customers online?

Aug 22, 2026 · 14 min read

Your customers are findable rather than guessable. Trace where the problem is already being complained about, separate presence from buying intent, and confirm one channel properly.

Your customers are online wherever they already complain about the problem you solve, and that is a findable place rather than a guess. Type the sentence a frustrated customer would type into a search engine and read what comes back: subreddits, trade forums, Facebook groups, review pages, comment sections. That is your shortlist. Then confirm one of those places with a small fixed test before committing months, because presence and buying intent are different facts. A platform where your customers exist is not automatically a platform where they buy.

Key takeaways

The question has a researchable answer. Guessing which platform to be on is optional.
Trace the complaint, not the demographic. People describe problems in public before they buy solutions.
Presence and buying intent are different. Being where they scroll is not the same as being where they hire.
One platform run properly beats six run badly, because every platform rewards frequency.
Search channels convert better; interruption channels teach you faster and cheaper.
Confirm a channel with a fixed test and a decided threshold before giving it a quarter of your life.

Most owners answer this question by picking the platform they personally use, or the one someone at a networking breakfast recommended. A forty-eight-year-old plumber ends up on TikTok because a video went around about how TikTok is the future. A bookkeeper who serves construction firms starts posting design-forward carousels on Instagram. Both are working hard on a channel selected by accident, and both will conclude in three months that online marketing does not work for their kind of business.

The alternative is not more sophisticated. It is just backwards from how most people do it. Instead of starting with platforms and asking whether your customers are there, start with the problem you solve and follow it to wherever people are already talking about it. That trail exists in public, it costs nothing to read, and it produces a shortlist of two or three places rather than a list of everywhere.

This question has an answer, and it is not a matter of taste

People describe their problems in public constantly and in enormous detail. Someone whose bookkeeping is nine months behind has posted about it. Someone whose lawn has been destroyed by grubs has asked about it with photographs. Someone deciding whether to hire a designer or use a template has started a thread asking exactly that, and forty people have argued underneath it. None of that is hidden and none of it requires a research budget.

What makes the trail useful is that it is written before anyone has sold them anything. The language is unfiltered, the anxieties are stated plainly, and the alternatives they are weighing are listed in their own words. That is more valuable than any platform demographic report, because it tells you not only where the people are but what to say when you get there.

The reason owners skip this step is that it feels like procrastination. Reading forum threads does not look like marketing. But an afternoon spent reading the actual complaints will change the channel decision, the ad copy, and often the offer itself, and it is the cheapest afternoon available in the entire business.

Start where the problem is already being complained about

The method is mechanical. Write down the sentence a frustrated customer would type at eleven at night, in their words rather than yours. Not "commercial landscaping services" but "grass dying in patches after last summer" or "how much should a bookkeeper cost for a small contractor". Then search that phrase and read the first two pages of results as a map rather than as answers.

Where the complaints actually live

Subreddits and trade forums. The highest-signal source for anything technical or trade-adjacent. People arrive to ask a specific question and stay to argue, which means the objections are visible too.
Facebook groups, especially local and occupational ones. Neighbourhood groups and "tradespeople in the county" groups are where local recommendations actually happen. Less indexed by search, so you often have to join to see them.
Review sections of the tools and services they tolerate. One-star and three-star reviews of your competitors and of adjacent products are a free list of unmet needs stated by people who already paid to have the problem solved.
Comment sections under how-to content. Under a YouTube video or blog post explaining how to do the thing yourself, you will find people explaining why the do-it-yourself version failed. Those are buyers.
Your own inbox and phone log. The questions you get asked repeatedly before someone books are the questions the whole market is asking. Most owners have never written them down.

What to write down when you find them

Keep the pass narrow and structured or you will read for three hours and retain nothing. For each place you find, record four things: the exact phrases people use for the problem, the alternatives they mention having tried, the objections they raise to paying someone, and whether anyone in the thread is asking for a recommendation. That last one is the difference between a place where the problem is discussed and a place where the buying happens.

Then talk to five existing customers, if you have any, and ask what they searched or where they looked before they found you. The answers are usually mundane and occasionally reveal a channel you had never considered — a supplier’s referral list, a local association directory, a single Facebook group with two thousand members that produces half your work. How to talk to customers before you build covers how to ask that without leading them.

Presence and buying intent are different facts

This is where most channel decisions go wrong. It is entirely possible for a platform to contain your customers and be a terrible place to reach them. A demographic report will tell you that plenty of small business owners use Instagram. It will not tell you that they use it in the evening, to look at holidays, in a state of mind that has nothing to do with hiring a bookkeeper.

A platform where your customers exist is not the same as a platform where your customers buy. The first is a demographic fact. The second is a habit.

Intent is visible in the wording of what people post, and it is worth grading. Someone writing "can anyone recommend a decent bookkeeper in the county" is in-market today. Someone writing "I hate tax season" has the problem and no intention of solving it this week. Someone liking a post about hating tax season has neither. All three are the same demographic and only one of them is a customer.

The practical consequence is that a small channel full of intent beats a large one without it. A trade forum with four hundred active members where three people a week ask for a recommendation is worth more to a service business than a platform with two million matching users who are scrolling for entertainment.

Search channels and interruption channels behave differently

Every channel available to you falls into one of two categories, and they are good at different jobs. On a search channel the person arrives already looking. On an interruption channel you appear in a feed and have to manufacture the interest yourself. Confusing the two produces the two classic errors: expecting feed advertising to convert like search, and expecting search to produce volume it does not have.

How channel types actually behave
Channel type	Intent when you arrive	Cost and speed of feedback
Search and map results	Someone is looking for the thing right now. The highest intent available anywhere.	Most expensive per click and worth it. Feedback in days if the term has volume, months if it is rare.
Paid social feeds	None. You are interrupting. All intent has to be created by the creative itself.	Cheap per click, expensive per customer. Readable in three to five days, which makes it the best testing surface you have.
Forums, subreddits, and groups	Mixed and legible. Some people are asking for a recommendation, most are venting. The wording tells you which.	Free apart from your time. Feedback on language in hours, on whether it produces work in weeks.
Short video and organic feeds	None, and the attention is borrowed. Strong for visibly demonstrable work, weak for abstract services.	Free to post, erratic to read. One post can outperform thirty for reasons you cannot reproduce.
Local listings and referral partners	High and geographically bounded. The person has already decided to hire somebody nearby.	Cheap and slow. Feedback in weeks to months, so it is a poor first test and an excellent long position.

The right sequence for most small businesses is to use an interruption channel to find out what language works, then apply that language to a search channel where the intent already exists. Paid social is a laboratory. Search is a harvest. How to choose a marketing channel for a small business works through that sequencing, and how to choose a channel for your first test covers the narrower question of which one to test against first.

Six platforms badly loses to one platform properly

Every platform is built to reward frequency. Posting four times in a week teaches the system who to show you to; posting four times in a quarter teaches it nothing, and the account decays. Attention is the constrained resource in a small business, and dividing it six ways guarantees that no single account ever crosses the threshold where distribution starts to compound.

There is a second, quieter cost. An abandoned account is worse than no account. A prospective customer who checks your Instagram and finds three posts from fourteen months ago has learned something specific and unhelpful about whether you are still in business. Six half-built profiles broadcast neglect five more times than one would.

The discipline is to pick the single highest-intent place from your shortlist, run it properly for ninety days, and only add a second when the first one runs without your daily attention. That threshold matters — a second channel added while the first still depends on you means both get done badly. How to post on social media consistently for a small business is about surviving those ninety days, which is where most attempts actually fail.

Local customers are a different problem from online ones

If your market is bounded by driving distance, the whole question changes shape. The total population that could ever buy from you might be forty thousand people, of whom a few hundred need the work this month. Broad platforms are structurally inefficient against a market that small, and you exhaust the reachable audience quickly, at which point the same people see your ad repeatedly and stop noticing it.

Local businesses get a compensating advantage: the highest-intent channel is also the cheapest to hold. A map listing with current hours, real photographs, and recent reviews captures people who are looking right now with no ongoing spend. Most local businesses treat that as an administrative chore and treat social media as the marketing, which is exactly inverted.

Where to start, by the shape of your customer
If your customers are	Start here	What confirms it
Within driving distance of you	Map listing and local search, then neighbourhood and county groups.	Calls that mention finding you on a map, and reviews arriving without being chased.
Anywhere, but in one occupation	The trade forum or subreddit for that occupation, plus occupational targeting on paid social.	Replies that use the trade’s own vocabulary and ask about your rate.
Anywhere, defined by a hobby or an identity	The community that already exists around it, then short video showing the work.	Saves and direct messages asking whether you take commissions.
Other businesses buying on an invoice	Search, plus one professional network and direct outreach to referral partners.	Enquiries that arrive already knowing your price range and asking about timelines.

The practical detail for the first row — listings, reviews, service-area targeting, and what to do when the whole market is four postcodes — is in how to market a local business online.

How to confirm a channel before committing months to it

A shortlist is a hypothesis. Confirming one takes about two weeks and a decided threshold, and the point of writing the threshold down first is that it stops you from reading whatever happened as encouragement.

Pick one place from the shortlist, the one with the most visible asking. Not the largest. The one where people request recommendations most often.
Write the threshold before you start. For example: three unprompted enquiries from people not already connected to you, within fourteen days.
Show up ten times in a way that is useful without you selling anything. Answer the question in the thread. Post the before and after. Explain what the do-it-yourself version gets wrong. Ten touches, not one.
Say what you do plainly once per touch, at the end. A single sentence identifying the work and the area. Concealing it wastes the test; leading with it gets you removed.
Log every response by type. Likes, replies, questions about the work, and enquiries. Only the last two count toward the threshold.
Compare against the threshold on the date you set, not on a good afternoon. Deciding early, in either direction, converts a test back into a guess.
If it cleared, run the same channel for ninety days before adding another. If it did not, move to the next place on the shortlist and repeat with the same threshold.

Channel confirmation: Evidence from a fixed, small test with a threshold decided in advance that a specific channel produces strangers who ask about the work — as opposed to six months of posting and inferring from the absence of results.

What confirmation looks like, and what it does not

The signals that matter are the ones that cost the other person something. Attention is free to give and therefore proves nothing. A question about availability costs a minute and some social exposure, which is why it means considerably more than a thousand views.

Counts as confirmation. Unprompted questions about price, availability, or scope. Direct messages from people with no prior connection to you. Someone else recommending you in a thread you are not in.
Weak but not worthless. Saves, follows from people who match the audience, and replies that add detail to the problem. These indicate the language landed even if nobody is buying this week.
Counts for nothing. Views, impressions, likes from other people in your trade, encouragement from friends, and a viral post that produced no enquiries at all.
A false negative worth checking. A channel that produced real interest and no enquiries usually has a conversion problem rather than a channel problem — no clear way to contact you, or no obvious next step.

If the interest was real and the enquiries were not, the failure is downstream of the channel and the fix is on your page rather than in your posting. Traffic but no sales deals with that case, and it is worth ruling out before abandoning a channel that was actually working.

When the honest answer is that they are not really online

Some customer groups are not gathered anywhere findable. Buyers of industrial equipment, owners of ageing housing stock, certain trade purchasers, and most people over seventy conduct their commercial lives at suppliers, trade counters, associations, and job sites. You can spend a year looking for their subreddit and it will not exist.

That is a legitimate finding rather than a failure of research, and it points somewhere specific. Those buyers still look you up before they call, so the online requirement is a search presence and a page that reads as real. The acquisition itself happens through referral partners, the counter staff at the supply house, and the association newsletter. The internet is a credibility layer rather than a channel, which is a much cheaper thing to maintain.

Once the channel is confirmed, showing up becomes a routine problem

The channel question is genuinely answerable in a fortnight. The part that defeats people is the two hundred days afterwards, because being present in a community is a daily obligation that competes directly with doing the work you are paid for. Almost every abandoned account belongs to someone who found the right place and then had a busy quarter.

This is the part Microphone automates, and it is worth being precise about how. After a claim has survived a real ad test, Microphone deploys social media subagents that comment, post, and like across Reddit, Instagram, X, TikTok, and LinkedIn on routines you set. They run on a Raspberry Pi 5 you own and keep in your house rather than on a rented cloud browser, which is the argument made in your social agent should live in your house.

The reason that matters for this question is sequencing. Automating presence on an unconfirmed channel just industrializes a guess. Confirm the place first with the two-week test above, then hand the routine over. How to run social media from a Raspberry Pi covers the hardware side, and how to find your first audience on social covers the first ninety days of actually being in a community rather than broadcasting at one.

One last note on cost. If you are choosing between channels partly on price, the comparison only makes sense per customer rather than per click, and the ceiling on what any channel can charge you is set by your own margin. What is a good cost per lead for a small business has that arithmetic.

How to choose a marketing channel for a small business — the decision framework in more detail, with sequencing
How to get customers from Reddit — the highest-signal channel for most trades, and how not to get banned
How to market a local business online — when your entire market is within driving distance
How to automate social media for a small business — what to hand over once a channel is confirmed
What is a good cost per lead for a small business — the arithmetic that decides which channels you can afford

Frequently asked questions

Where are my customers online?: Wherever they already complain about the problem you solve, which is a findable place rather than a guess. Search the exact phrases a frustrated customer would type, and see which forums, subreddits, groups, comment sections, and review pages come back. Those results are your shortlist. Then confirm one of them with a small fixed test — a post, a reply, or a narrow paid campaign — before committing months to it. Presence is not the same as buying intent, so the confirmation step is what turns a plausible platform into a real channel.
How do I find where my target audience spends time?: Work backwards from the complaint rather than forwards from the demographic. Type the sentence a frustrated customer would type into a search engine and read what ranks: subreddits, trade forums, Facebook groups, review pages, YouTube comment sections. Then ask five existing customers where they looked before they found you. Those two lists overlap more than most owners expect, and the overlap is where to start.
Should my small business be on every social platform?: No. One platform posted on consistently beats six posted on occasionally, because every platform rewards frequency and punishes absence. Spreading a limited amount of attention across six accounts produces six dead accounts, each of which makes the business look abandoned to anyone who checks. Pick one, run it properly for ninety days, and add a second only when the first runs without your daily involvement.
What is the difference between a search channel and an interruption channel?: On a search channel the person arrives already looking for what you sell, so intent is high, competition is priced in, and volume is capped by how many people are searching. On an interruption channel you appear in a feed and have to create the intent yourself, so volume is nearly unlimited and quality is lower. Search converts better. Interruption tests faster and cheaper, which makes it the better place to learn what language works.
How do I know if a channel is working before I have spent months on it?: Decide the threshold in advance and run a fixed two-week test. Look for people who were not already connected to you asking a question about the work — not likes, not follows, not views. Three unprompted enquiries from strangers in two weeks is a real signal at small scale. Zero, from a channel where you posted honestly and specifically ten times, is also real information.
Where do local businesses find customers online?: Local businesses have a narrower and easier answer: map listings and local search first, neighbourhood groups second, and referral partners third. The population that can ever buy from you is bounded by driving distance, which makes broad platforms inefficient and makes reviews and local search unusually powerful. A well-maintained map listing with recent reviews often outperforms every social account combined.
What if my customers are not really online in a findable way?: Some are not. Buyers of industrial equipment, elderly homeowners, and certain trade purchasers gather at suppliers, trade counters, associations, and job sites rather than in public forums. In that case the online channel to use is search, because they will still look you up before calling, and the acquisition happens through referral partners and physical presence. That is a legitimate answer, not a failure to find them.
How long should I give a channel before deciding it does not work?: Ninety days of genuine consistency for an organic channel, and two to three weeks for a paid one. The difference is that paid buys attention immediately while organic has to accumulate it. What ruins most organic tests is not the ninety days but the seven-post start followed by a three-week silence, which measures your consistency rather than the channel.