Run paid ads and scale lead flow 2 of 23 in this group
SOP 36
Build the targeting ladder
What this page is for. Use it to decide who sees the advertisement you paid for. The ladder runs from everybody alive down to a named group, and every rung narrows it. This page gives the two targeting methods, the order to apply them in, the exception for local businesses, and the action steps for assembling the lists the methods run on.
SOP-36-Build-the-targeting-ladder.md
1. Why this is the problem worth solving
Targeting is one of the hardest parts of paid advertising, and the view here is that it is the only problem to solve.
The reasoning is worth following, because it changes what you work on. Put the right people in front of an exceptional lead magnet or offer and they engage. When paid advertising is not efficient enough, then, what that tells you is that too few of the right people are seeing it — and the work is the targeting rather than the copy.
That reduction is the payoff of this page: a paid-advertising problem is a targeting problem, which simplifies a great deal what there is to solve.
Flag: targeting, or the ad itself? A second view, from a paid-ads team, concerns one large ad platform as it works now: the platform has become very good at finding buyers inside almost any audience it is given, so the ad itself tells it whom to find. That team never targets by interest; audiences of people who engaged with the business, and look-alikes, still perform for it, but broad targeting is its default. An ads practitioner adds that how an account is set up matters far less than people think. The same flag covers section 4's interest targeting, and sits on SOP 213, section 5. This page does not settle which reading is right.
That hooks decide who comes in is SOP 212, section 4; judging single ads against a target cost is SOP 275.
2. The ladder
You start with everybody and narrow, rung by rung.
| Rung | What it removes |
|---|---|
| 1 | Choosing a platform removes most of the world in one step |
| 2 | A look-alike audience, grown from a list you supply |
| 3 | Filters — age, income, gender, interest, time, location |
The first rung, in figures. Pick a platform with 100 million users and you have removed 99 percent of the world before any targeting. If all your buyers share one language and half the platform's users do not, excluding them takes the cut to 99.5 percent. Choosing a platform is itself the first and largest cut, removing the great majority of the world before you have targeted anything.
Be too specific rather than too broad. The same widening rule that governs subject matter on SOP 25 governs audiences: start with the smallest audience, because that is where the return is highest; the return then affords you the money to be a little less efficient as you widen; and that affords more again, until — if you are good enough — you can market to the whole world.
3. Method one: the look-alike audience
A modern platform will take a list you hand it and put your advertisement in front of a far larger audience that resembles the people on that list.
The list is the input, and its quality sets the ceiling.
| The list you upload | What the platform finds you |
|---|---|
| A thousand actual customers | More people like your customers |
| People who merely opted in | More people who merely opt in — not as good as customers, better than a cold audience |
Bigger and better both improve the result. The lists you already have are the warm-outreach list from SOP 17 and the cold list from SOP 27, and you add them together if you need the size.
There is a minimum list size before a platform will build a look-alike from it. This page gives no figure for that minimum.
4. Method two: filters
Target on a factor of your choosing: age, income, gender, interest, time, geography.
| If | Then |
|---|---|
| Nobody over 45 or under 25 is in your market | Exclude them |
| You sell car parts | Show the advertisement during car programming and on car channels |
| Only people with pets buy | Include pets as an interest |
The two methods combine. Build the look-alike from your list, then add age, gender, income, interest, time and geography on top, and the audience narrows again. You could use one, or the other, or both — it depends on the size of the pie you are looking at.
If you cannot build a look-alike audience at all, because you have no usable list yet, start by targeting interests on their own — people with pets, in an area, in a state.
5. The local exception
Local markets are already tiny beside national ones — a local market on its own is around one tenth of one percent of a nation — so the ladder stops early.
| Situation | Rule |
|---|---|
| Marketing to a two-to-three-mile radius | Do not add more than maybe one filter — maybe older people, maybe men, maybe women, and not much beyond that |
| A very small market | You will reach everybody for a fairly nominal spend, so some inefficiency is acceptable |
Be as specific as possible, but no further. Adding filters to an audience that is already narrow removes people you needed rather than people you were wasting money on.
Local advertising is also one of the easiest kinds to run, and the reason is trust rather than technique: you already have it. Naming a landmark people know — "we are next to the place on Fifth" — produces instant recognition, which means you do not have to sell anything in the advertisement at all. You get them to opt in, and then you call them. One of the easiest categories of advertising to run is a local one, and a local business owner with the offer right can usually just target the area and be fine.
6. The action steps
| Step | What you do |
|---|---|
| 1 | Bring every lead list you have into one place |
| 2 | Separate them into past and present customers, warm outreach, and cold outreach |
| 3 | If any one of those lists is big enough, use the ones closest to your customers |
| 4 | If none is big enough on its own, combine them into one list |
| 5 | Over time, as you accumulate customers or booked calls, sift those out, re-upload them, and build a look-alike from that list |
| 6 | Add your targeting |
| 7 | If the platform allows it, use the lists in order of quality to build the look-alike audience |
| 8 | Add your filters on top |
7. The checklist
| Question | The answer |
|---|---|
| Who runs this | You |
| What you are narrowing | Everybody alive, down to a named group |
| The first cut | Choosing the platform |
| The two methods | A look-alike audience, and filters |
| What a look-alike needs | A list you upload |
| What sets its quality | The quality of the list — customers beat opt-ins, opt-ins beat cold |
| Which lists to use | Customers, then warm outreach, then cold outreach |
| The filters available | Age, income, gender, interest, time, geography |
| The local rule | At most about one filter |
| What to do with no list | Target interests on their own |
| The minimum list size for a look-alike | Not established on this page |
| How narrow is too narrow | Not established on this page |
8. What this page does not cover
Choosing the platform in the first place is on SOP 35. The wording and the imagery of the advertisement itself sit on SOP 32, SOP 33, SOP 34 and SOP 39; SOP 31 sets the reading level for every part of it. SOP 37 covers the page people arrive at. The lists this targeting runs on are built on SOP 17 for warm contacts and SOP 27 for cold ones. Widening the subject you talk about follows the same smallest-to-largest shape, and is on SOP 25.
Bidding, budgets, audience overlap and the cost arithmetic that decides when to widen are not covered on this page.
Terms defined on this page
- Look-alike audience
- A larger audience a platform builds to resemble a list you upload, such as buyers, calls or opt-ins, in sizes of 1, 2 and 3 percent. Its quality depends on the list: customers beat opt-ins, and opt-ins beat cold contacts.
- Targeting ladder
- Narrowing an ad's audience rung by rung: pick the platform, then a look-alike audience, then filters. Err toward too specific rather than too broad; for a tiny local radius, add perhaps one filter at most.