What's the best Facebook ads scaling strategy for lead generation?

Scale two ways at once: raise the budget on a winning ad where it sits, by no more than about 20% a day, and keep launching new ads. That way the next winner is always on its way. Never move a winner into a new "scaling campaign" and never edit it.

Most scaling advice treats this as a campaign structure problem. From what I've seen, it's an ad supply problem. You can only scale as far as your ads can carry you.

What most people get wrong about scaling

Someone finds an ad that's bringing in booked appointments. They get excited. They build a new campaign, move the winner across, triple the budget and change the audience while they're at it. A week later the ad that was working isn't.

Then they blame the algorithm. The algorithm didn't break the ad. They did, by changing everything that was working at the same time.

First, know what a winner is

You can't scale an ad until you know it's a winner, and one good day doesn't tell you that. I never judge an ad daily.

At the four-day mark I check a new ad's cost per lead against a kill line. Cost per lead is the fastest number I've got, so it's an early warning. If the ad is over the line, it's off. If it passes, it runs to around 20 leads. Then I judge it on how many of those leads became qualified, booked appointments.

Cost per booked appointment is my north star. It's the number that tells me whether the advertising is working. An ad with cheap leads that never book isn't a winner. An ad with dearer leads that turn into booked appointments might be.

Why not judge each ad on sales? Sales take too long and there are too few of them per ad. Qvolt's sales cycle is about six weeks. You'd wait six weeks to judge each ad, on maybe one sale each. That's not statistically significant. Selling is the sales team's job. The ad's job is to put booked appointments in front of them.

What a customer is worth still matters. It's what tells you how much you can afford to pay for an appointment. A winner is an ad bringing in booked appointments for less than that.

If you don't track leads through to booked appointments, you can't scale with any confidence. You'd be pouring more money onto a number that stops at the form.

Vertical scaling: raise the budget slowly, where the ad sits

Vertical scaling means giving a winner more budget. The mistake is doing it in one jump.

I change the budget on a winner by no more than about 20% in any 24 hours. Up or down. The ad stays in the same ad set, in the same campaign, with the same audience. Nothing else changes.

Here's what that looked like on Qvolt Solar, the solar business I co-own. One winning ad was holding $259 AUD per customer at $50 AUD a day. That's what each customer cost once the sales came through, and it shows the ad was good all the way down the line. For one ad recently, I raised its budget 20% every 24 hours until its cost started to climb. Then I stopped pushing that ad and scaled sideways instead, with new ads.

So i prefer to find the ad's ceiling if you will, before leaving the daily spend where it is - then launching additional ads, with the ad spend i would've put on the original ad.

From what I've seen, every ad has a ceiling, and some ads hold their cost per booked appointment much further up than others. You won't know where the ceiling is until you find it. That's why the steps are small.

Horizontal scaling: keep launching new ads

Horizontal scaling means more winners, not a bigger budget on one. This is the part people skip, because it's hard work. A good ad takes real thought to write and make.

No single ad carries a campaign forever. If you stop launching because you've found a winner, you're scaling on borrowed time. I explain why I keep launching new ads even when the current ones are working in this video.

Each new ad goes into its own ad set at about $50 a day [REV: confirm currency] and gets the same test as every other ad: cost per lead at four days, then booked appointments once it's had around 20 leads. The ones that prove themselves become the next winners you scale vertically.

Make the new ads meaningfully different. Different ad types, different angles, different bait. Not the same ad with a new headline.

Why you never restructure a winner

The "scaling campaign" is the most common way I see people kill a good ad. They pull their winners out of where they were proven and pack them into one new campaign with a big budget.

Leave winners where they sit. If you want a version of a winner somewhere else, or with a change, duplicate it. Never edit the original. An edit throws away what the ad has learned and turns a proven ad back into a test. I answer whether you should move winners into one ad set in this video.

The same goes for daily fiddling. Budget up, budget down, new audience, new headline. Every small change resets what the ad has learned. The tinkering is the problem.

Do both, when the budget allows

Vertical and horizontal aren't a choice. Raise budgets slowly on the winners you have, and keep new ads being tested beside them.

Your budget decides how many new ads you can test at once. New ads come first. Rather than putting more on one winner, I use that money to launch and prove a new ad, until I've got somewhere around 8 to 10 proven ads. That way, if one or two drop off, lead flow doesn't go to zero overnight.

And don't add awareness campaigns as part of "scaling". For lead gen, stay on lead form or conversion objectives. Brand retargeting only comes in once lead gen is already profitable.

FAQ

How much should I increase my Facebook ad budget when scaling?

No more than about 20% in any 24 hours, on an ad that has already proven itself on cost per booked appointment. Big jumps are how people break the ads that were working.

What's the difference between vertical and horizontal scaling?

Vertical scaling is raising the budget on a winning ad where it sits. Horizontal scaling is launching and proving new ads so you have more winners. I do both when the budget allows.

Should I move my winning ads into a new scaling campaign?

No. Leave winners where they were proven. If you want a variation, duplicate the ad rather than editing it, and keep launching new ads alongside it.

Will raising my budget reset the learning phase on a winning ad?

From what I've seen, it's big jumps and constant edits that break a winning ad. I raise a winner by no more than about 20% in any 24 hours, where it sits, and change nothing else. Beyond that, I don't worry about the learning phase.

When should I stop raising the budget on an ad?

When its cost per booked appointment starts to climb. That's the ad's ceiling. From there, scale horizontally by launching new ads.