How to Test 5 Job Titles for Under €200

Something that keeps bugging me: a company can spend weeks on the text of a job ad, and then the job title gets picked in about ten seconds. Usually it's just whatever the role is called internally. But the title is the first thing anyone reads, and in a feed it's often the only thing they read. So why are we guessing?

You can just test it. Five titles, the exact same ad, the same budget for each, and after a week you know a lot more about which title gets people to actually finish an application. Not proof, I'll be honest about that further down. But a direction, and a direction beats a gut feeling.

What you need

• A Meta ad account with a payment method, and access to the Facebook page you advertise from.
• A way to count finished applications per ad: a Meta instant form, or tracking on your "thanks for applying" page so Meta can see who finished. Without this you only see clicks, and clicks don't fill jobs.
• One image or video without the job title in it.
• Five job titles for the same job.
• Up to €200, 7 days and a spreadsheet.

Step 1: Pick five titles worth testing

Don't test five spellings of the same thing. "Sales Manager" vs. "Sales manager (m/f)" tells you nothing. You want five actually different ways a real person could describe this job. Here's a made-up example for a planning role at a transport company:

One rule: every title has to describe the same job honestly. If one makes it sound like a different job, you're testing a different vacancy, not a title. And no "Logistics Ninja". That's check 1 of the Job Ad Scorecard for a reason.

Step 2: Build one ad and freeze everything else

This is where most tests break. If anything else differs, you don't know what made the difference. So you build one ad, and only the title changes.

Headline:
{TITLE} in {city} | {salary range}

Primary text:
We're hiring a {TITLE} in {city}.
{One sentence on what you actually do all day.}
{One sentence on what's in it for them: salary, hours, team.}
Applying takes about {X} minutes, no cover letter needed.

Button: Apply now
Image or video: the same file for all five, no title in it.

Copy it five times and only swap {TITLE}. If your image has the title written on it, make one without, otherwise you're changing two things at once.

Step 3: Set up the campaign in the employment category

When you create the campaign, you declare it as a Special Ad Category: Employment. According to Meta's developer documentation on special ad categories, that's required if you're a US advertiser or your ads reach people in the US, Canada or Europe. The same page lists what changes:

Meta changes this stuff now and then, so check the page before you start. But here's what I find interesting. Because you can't target narrowly, the title does a big part of the filtering for you. The right person has to read it and think "that's me." So the title is basically your targeting now.

Give all five versions the same location, the same placements and the same goal: leads or finished applications, not clicks.

Step 4: Split the budget equally across the five titles

Two ways to do this.

Option A: Meta's A/B test feature. Meta's A/B testing page says it splits audiences randomly so the groups are comparable, that you should change only one variable, and that you can test up to five variants. Five titles, one variable, perfect fit. The button moves around in Ads Manager sometimes, so if you can't find it, search "A/B test" in Meta's Business Help Center.

Option B: five ad sets by hand. One campaign, five ad sets, one ad each, and every ad set gets a lifetime budget of €40. Put the budget on the ad sets, not on the campaign. Otherwise Meta decides how to divide the money, and one title can end up with most of it. A bit less clean, since one person can see more than one version, but still useful.

Either way it's €40 per title, a bit under €6 a day. Small, I know. That's why step 6 matters so much.

Step 5: Run it for 7 days and measure cost per finished application

Seven days, so every version gets weekdays and a weekend. And don't touch it. No turning off the "bad" one after a day, no tweaking the text. Only fix things that are actually broken, like a rejected ad or a form that doesn't send.

Then fill in this table:

The number that matters is cost per finished application, and right next to it, how many of those people you'd actually call. A clickbaity title can get lots of clicks from people who don't fit. So go through the applications per title and just mark them yes or no. That last column often tells a different story than the clicks.

Step 6: Decide if it's real or just luck

Okay, the honest part. With €40 per title you'll be working with small numbers. And small numbers jump around a lot, just by chance.

Here's a quick way to feel how much. Take the square root of the number of applications a title got. That's roughly how much that number could move if you ran the exact same ad again next week.

• Title A got 9 applications. The square root is 3, so next week it could easily be 6 or 12.
• Title B got 6. The square root is about 2.5, so it could easily be 4 or 8.

Those ranges overlap a lot. So A "beating" B is probably luck.

My rules of thumb, and they're rules of thumb, not a statistics test:

• One title gets zero to two applications while the others get around ten? Take that seriously. Flops are easier to spot than winners.
• The top two are close? Call it a tie. Pick the clearer one, or test just those two later with more budget.
• One title has roughly double the next one, on the same spend, and both are more than a handful? Now it starts to look like a real direction.

If you used Meta's A/B test feature, it can also show how confident it is about the result. If it says confidence is low, believe it.

So what €200 gets you is a direction: which titles to drop, and which one or two deserve the next round.

What "done" looks like

• One employment campaign with five versions that only differ in the title.
• Each version spent about €40 over 7 days, without changes halfway.
• A filled-in results table, including the "would I call them" column.
• A short note: which titles flopped, which look strongest, and whether the gap is big enough to trust.
• The winning title live on your job ad, careers page and job boards too, not only in the Meta ad.

Common mistakes

The title is check 1 of the Job Ad Scorecard. Once you have one that works, the next question is where people drop off after they click, and that's what the Hiring Funnel Leak Map is for.

Sources

• Meta for Developers: Special Ad Categories
• Meta for Business: A/B testing