Target your social media ads to the people who'll actually buy — four steps that narrow the field

Targeting social media ads means deciding who sees your paid posts before you pay for them. You set the criteria—age, location, interests, behaviour—and the platform shows your ad only to people who match. The tighter your aim, the less you spend on people who won't buy. HW Media Management walks you through the mechanics below: what the options mean, how to combine them and when to stop narrowing.

Start with what you already know about your customers

Before you open the ad manager, write down who buys from you now. Age range, where they live, what they searched for before they found you. If you sell yoga mats, your buyers might be women aged twenty-five to forty in cities, interested in fitness and mindfulness. If you fix commercial boilers, your audience is facility managers in a fifty-mile radius who searched "boiler repair near me".

This is your starting profile. It stops you guessing and gives the platform something concrete to work with. The profile doesn't need to be perfect—you'll refine it once the ads run—but it needs to reflect real customers, not a hunch about who you'd like to reach.

Pick your targeting layer: demographics, interests or behaviour

Social media platforms let you target by three main layers, and each one filters your audience differently.

Demographics narrow by who people are

Age, gender, location, language, education, job title. These are the simplest filters and the ones you'll use on every campaign. A restaurant in Manchester targets people within ten miles; a recruiter for graduate schemes targets eighteen- to twenty-two-year-olds studying at UK universities.

Interests narrow by what people follow

Platforms infer interests from the pages someone likes, the posts they engage with and the groups they join. If your product is outdoor gear, you can target people interested in hiking, camping or trail running. Interests work well when your product aligns with a clear hobby or lifestyle, but they're less reliable for niche services.

Behaviour narrows by what people do

Purchase history, device usage, travel patterns, life events. Behavioural targeting is the sharpest layer because it's based on actions, not stated preferences. You can reach people who've bought online in the past month, who commute by train or who've recently moved house. Not every platform offers the same behavioural data, so check what's available in your ad manager.

Combine layers without stacking too many filters

You can layer demographics, interests and behaviour to shrink your audience, but every extra filter makes the pool smaller. Target women aged thirty to fifty in London who like yoga and bought fitness equipment online, and you might end up with an audience too small to reach. The platform needs enough people to show your ad to, or it won't spend your budget.

A useful rule: start with one or two demographic filters, add one interest or behaviour layer, then stop. Run the ad for a week and check who clicks. If the clicks come from the right people, leave the targeting alone. If they don't, adjust one filter at a time so you know what made the difference.

Use exclusions to cut out the wrong people

Exclusions are filters that keep certain groups out. If you sell premium products, exclude people who've clicked your discount codes page. If you're recruiting, exclude your own employees so they don't see your job ads. If you're promoting a local event, exclude anyone outside the region.

Exclusions stop you wasting impressions on people who won't convert, and they're faster to set up than you'd think. Most platforms let you exclude based on the same criteria you use to include: age, location, interests, behaviours. You can also exclude custom audiences—lists of people you upload yourself, such as existing customers or newsletter subscribers.

Test a lookalike audience when you've got a list that converts

A lookalike audience is a group of people the platform finds based on an uploaded list—your best customers, your email subscribers, people who bought in the past six months. The platform analyses that list, works out what those people have in common, then finds others who share the same traits. You can read how lookalike audiences work in more detail if the concept is new.

Lookalikes work best when your source list has at least a hundred people and they've all done the same thing: bought, signed up, attended. The more consistent the behaviour, the sharper the lookalike. You set the size—one per cent of a country's population is narrow and similar to your source; ten per cent is broad and cheaper to reach—but start narrow and widen only if you're not getting enough reach.

Watch the estimated audience size before you launch

Every ad manager shows an estimated reach: the number of people your targeting will let you show the ad to. Too small—under a thousand—and the platform struggles to spend your budget. Too large—millions—and your ad becomes background noise. The sweet spot depends on your budget and how long the campaign runs, but as a guide, aim for ten thousand to a hundred thousand for a local or niche campaign, and a hundred thousand to a million for a national one.

If the estimate is outside that range, adjust your filters. Widen demographics or drop an interest layer if the audience is too small; add an exclusion or tighten the location if it's too large. The estimate updates live, so you'll see the effect of each change immediately.

Set up your first campaign in four practical steps

1. Open your ad manager and create a new campaign. Choose your objective—awareness, traffic, conversions—then name the campaign something you'll recognise later.
2. Set your daily budget. Start low—ten to twenty pounds a day—so you can test without overspending. You can increase it once you know the targeting works.
3. Build your audience. Add demographics first, then one interest or behaviour layer. Check the estimated reach and adjust if it's too narrow or too broad. Add exclusions if you know groups to avoid.
4. Write your ad, upload your image or video, then launch. Let it run for at least three days before you judge the results—platforms need time to find the people who'll respond.

Refine your targeting after the first week

Most platforms show you a breakdown of who clicked, who converted and who ignored your ad. Look at age, gender and location first: if one group converts well and another doesn't, shift your budget toward the group that works. Then check interests and behaviours. If "yoga enthusiasts" clicked but didn't buy, drop that interest and try "fitness equipment buyers" instead.

Refining is faster than starting over. Keep the parts that work, replace the parts that don't, and give each change a few days to settle before you tweak again. Over-adjusting daily wastes time and confuses the platform's learning phase.

Compare platform targeting options before you commit

Not every platform offers the same filters, and knowing the differences saves you time.

Choose your platform based on where your audience spends time and what data you need to reach them. A graphic designer targeting creative directors uses LinkedIn's job-title filter; a jeweller targeting engaged couples uses Facebook's life-events filter.

Avoid over-targeting by keeping one variable loose

The tighter you squeeze, the fewer people see your ad, and platforms perform worst when the audience is too small to test variations. If you've set age, location, three interests and two behaviours, you've probably gone too far. Drop one layer—usually the least relevant interest—and let the platform find the right people within a broader pool.

Over-targeting also costs more. When the audience is tiny, competition for those impressions drives up the price per click. A slightly wider audience often delivers better results at a lower cost because the platform has room to optimise.

Track who converts, not just who clicks

Clicks tell you someone was interested; conversions tell you they bought, signed up or did what you asked. Set up conversion tracking in your ad manager so you know which audience segments actually deliver. If thirty-year-old women click but forty-year-old women convert, shift your budget and your creative toward the group that completes the action.

Conversion data also feeds back into lookalike audiences. The platform learns which traits predict a sale, and your next lookalike becomes sharper. Without tracking, you're guessing; with it, you're refining based on what worked. If you're unsure how to interpret the numbers, reading a social media analytics report explains what to look for.

Scale by duplicating what works, not by widening everything

When an ad performs well with one audience, duplicate the campaign and target a similar group—a neighbouring city, a related interest, a slightly older age bracket. Don't widen the original campaign to "see what happens"; you'll dilute the data and lose track of what's working. Duplication lets you test new segments without risking the performance of the one that's already delivering.

Each duplicate should change only one variable. If your original ad targets Manchester, duplicate it for Liverpool. If it targets yoga enthusiasts, duplicate it for pilates enthusiasts. Keep everything else—creative, budget, exclusions—the same, so you know whether the new segment works as well as the first.

#Audience targeting #Paid ads

Download PDF: Social media ad targeting checklist PDF, 30 KB