Facebook and Instagram ads in Newcastle, NSW: reach local buyers.

Search reaches people who typed the thing in, and this reaches people scrolling past it. The interest has to be created as well as captured, which is why the advertisement itself carries so much of the result.

That changes what an advertisement has to do. It has to earn the pause before it can make the offer at all, which is why the creative decides more of the outcome here than the targeting ever does.

What this channel suits

• Something people would never search for.

  A new service, an offer, an event. Where nobody is typing it, this is how they find out it exists at all.

• Something with a picture in it.

  Work that can actually be shown. A before and after, a finished job, a room, a face somebody recognises.

• A local audience that is worth reaching.

  A radius around Newcastle is straightforward to set, and a local business is recognisable in a way a national one is not.

• Reaching the same people a second time.

  Somebody who visited the website and did not enquire, shown something a second time a week later.

The platform will find the right people where the advertisement gives it something to work with. A weak advertisement targeted perfectly does less than a strong one aimed broadly.

Advertisers spend most of their attention on targeting because targeting has settings and the creative does not. The platform has spent years getting good at finding people, and it needs something worth showing them.

What actually moves it

• The offer sitting behind the advertisement.

  What somebody gets and why now. No amount of design rescues an offer that nobody actually wants.

• The first second of the advertisement.

  The picture, or the opening frame of the video. That is what decides whether the scroll stops at all.

• Whether the whole thing looks real.

  Your own work, your own premises, your own people. Stock photography reads as an advertisement instantly, and people scroll straight past it.

• The first line of the words.

  Written the way a person says it rather than the way a brochure does.

This is also the honest reason results vary so much between businesses running the same budget on the same platform. It is very rarely the settings.

Somewhere for the enquiry to land, somebody to answer it quickly, and a way of counting what arrived. Paid social produces enquiries fast, and it exposes whatever happens next.

This channel finds people who were not looking, which means they are less committed than somebody who searched for you. What happens in the hour after they enquire matters more here than anywhere else.

What has to exist first

• A destination asking for one thing.

  A page or a form built for the offer in the advertisement, written in the same words as the advertisement.

• Somebody available to answer it.

  Answering quickly is the point. An enquiry from a feed goes cold faster than one from a search, because the person was not looking in the first place.

• Counting that already works properly.

  What arrived, and from which advertisement. Otherwise the whole thing comes down to a feeling at the end of the month.

• Enough budget to actually learn something.

  The platform needs a run of results before it can steer. A few dollars a day for a week teaches nobody anything at all.

Meta sets out what it needs and what it will not allow in its advertising standards. Reading the rules for your industry before writing anything saves a rejected account later.

Boosting posts instead of running advertisements, changing everything at once, judging it in the first week, and sending clicks to a home page. All four are common and all four are avoidable.

These are the ones that turn up whenever a business says social advertising did not work for them.

What goes wrong

• Somebody boosts a post instead.

  The button on a post is a simpler product with far fewer controls. It is not the same thing as running an advertisement.

• Changing everything at once.

  A new picture, new words and a new audience in the same week. Nothing comes out of it, and the platform starts over.

• Judging it in the first week.

  The early days are the platform working out who responds. A verdict then is a verdict on noise.

• The clicks go to the home page.

  Somebody who was not looking for you arrives at a menu of options, and leaves without using any of them.

Settle the offer, gather real photographs, and build the destination. Put the counting in, run a small number of clearly different advertisements, and then change one thing at a time.

The first two steps cost nothing and decide most of the result.

1. Settle the offer

What somebody gets and why now. It decides more than everything else combined, and it is free to change right up until the spending starts.

2. Gather real material

Photographs and video of the business's own work, premises and people. An afternoon with a phone is usually more than enough for it.

3. Build the destination

A page or a form for that one offer, written in the same words as the advertisement that sends people to it.

4. The counting goes in

What arrived, and from which advertisement, in place before the first dollar gets spent on anything.

5. A few clearly different advertisements run

Different pictures and different opening lines, not three shades of the same one. The question is which idea works, not which font.

6. Then one thing changes at a time

With long enough between changes for the platform to steer, before anybody decides what worked.

Enquiries can arrive on the first day. Knowing whether it works takes weeks, because the platform spends the early period learning who responds to it.

The order it happens in

• First the offer and the material come together.

  Nothing is running at this point, and this is the part deciding the result more than any setting.

• Then it runs and it learns.

  The early period is the platform finding who responds. Results are uneven then, and that is entirely normal.

• Then it settles, and changes one thing at a time.

  New pictures go in as the old ones tire, which they reliably do.

Advertisements wear out on this channel in a way search ads do not, because the same people keep seeing them. New material is an ongoing part of the work rather than a one off.

Ask whose advertising account it runs in, who makes the creative, what is being counted, and what the fee is separately from the spend. The account question decides what you keep.

These are worth asking of any agency, including this one, so the answers given here are underneath each question.

• Whose ad account does it run in?

  The business's own, with the agency added to it. An account owned by an agency takes the history and the audiences away with it.

• Who is actually making the creative?

  It moves the result more than the targeting does, so it is worth knowing whether it is being made or recycled.

• What is being counted, exactly?

  Enquiries, and which advertisement produced each of them. Reach and engagement only describe activity.

• What is the fee, separately from the spend?

  Two separate numbers, both of them named. A percentage of spend is a fee growing every time the budget does.

Where people already search for what you sell, search work and search advertising are usually the cheaper starting point. This channel earns its place when the interest has to be created.