Is a brand lift study worth it on Meta or TikTok?
When a brand lift study on Meta or TikTok answers a real budget question, when it is only a nice slide, and why smaller accounts often cannot run one.
When a brand lift study on Meta or TikTok answers a real budget question, when it is only a nice slide, and why smaller accounts often cannot run one.
The client said it on the kickoff call: "This video push is about awareness, so don't judge it on ROAS." Six weeks later the same client opens the report, sees a last-click ROAS of 0.6 and asks what the money bought. You have no answer that isn't a feeling. That is the moment someone suggests a brand lift study, and it is the right moment to ask whether one would help.
Short answer: A brand lift study is worth running on Meta or TikTok when its result would change a budget decision, such as keeping an upper-funnel video line that last-click reporting cannot credit. The study compares poll answers on ad recall or awareness between exposed and held-out users, says nothing direct about sales, and platform access is gated.
The takeaways
A brand lift study measures the difference in survey answers between people who saw your ads and people who were held back from seeing them. TikTok's help article describes it as two randomized groups drawn from your target audience, an exposed group and a control group, with the platform comparing "key performance indicators, such as awareness and ad recall" between them. Meta's version works on the same holdout logic, with a short poll shown in feed.
So the output is a statement like "people who were eligible to see the campaign said yes to 'Do you recall seeing an ad for this brand?' more often than people who were not." That is a causal read, because the groups are randomized. It is a causal read on a survey answer, though. Nobody bought anything in that sentence. Whether recall turns into revenue next quarter is a separate assumption you still have to defend.
A brand lift study earns its cost when you can name the decision before it runs. Write one line: "If recall lift is above X points, we keep the video line at this budget. If it is flat, we move the money to the conversion campaigns." If you can't write that line, the study will produce a number and no action.
The cases where I've seen that line hold up are narrow. A client who has agreed in advance to judge an upper-funnel budget on something other than ROAS. A video or reach campaign whose effect lands in branded search or a later purchase, where the last-click report will always show close to zero. A creative choice between two very different brand stories, where you want to know which one people remember at all.
In each case the platform report cannot answer the question, and the decision is big enough that a better-than-gut answer matters.
A brand lift result turns into decoration when nobody agreed beforehand what a good or bad number would trigger. A positive recall lift lands in the quarterly deck, everyone nods, and the budget stays where it was. A flat result gets explained away as "awareness takes time." Either way the study changed nothing.
Two reading traps make this worse. First, percentage points versus relative lift: going from 20% to 22% recall is 2 points, or 10% relative, and a report can lead with whichever sounds better. Ask for both and the baseline. Second, ad recall is a weak proxy. People misremember which brand they saw and where, which is why recall reads best as a comparison between your own campaigns (did the new creative beat the old one on the same question?) and worst as an absolute proof of value.
If the honest plan is "we'll see what it says," skip the study and keep the budget in the test you can read.
Often the first barrier is access. TikTok's help center states that Brand Lift studies "can only be set up by your account manager," which means an account without a managed rep has no self-serve path. Meta's developer docs describe Conversion Lift access as "currently limited" and send you to your Meta representative. I checked the TikTok article and Meta's lift-study developer guide for a published minimum spend and found none, so treat any fixed dollar figure from a third-party blog as hearsay until your rep confirms it for your market.
The structural reason for a gate is sample size. A poll only reaches a slice of each group, and a gap of a few points needs many answers on both sides before it stops being noise. Small campaigns produce wide uncertainty, so the platforms keep the study for budgets that can feed it. If you are below that line, a vague result is the best case.
No, and mixing them up is the most common way a brand lift study gets oversold. Brand lift asks people a question and compares their answers. Conversion lift compares what people did, purchases or sign-ups recorded by your pixel or events, between exposed and held-out groups. Both use a randomized control. Only one of them touches revenue.
If the client's real question is "would these sales have happened without the ads?", that is an incrementality question, and I've written up the cheaper holdout designs in incrementality testing for Facebook ads. If the question is "do people remember us after this campaign?", brand lift is the right tool, assuming you can get access. Picking the study that answers the client's sentence saves you from presenting a recall number to someone who asked about sales.
Below the gate, your own ad-level data still answers the creative half of the question: which video holds attention, which hook stops the scroll, which ad is fading. That is the part I built Adscalr around. It does not run lift studies or surveys. It reads ad performance from connected Meta and TikTok accounts, scores each ad on six metrics including hook rate and CTR, and pulls a new ad's early score toward what its format usually does, so one lucky week doesn't pass for a winner. The scoring method is on the ad intelligence page. It won't tell you whether people remember the brand, and nothing in an ads dashboard can.
This is the thinking behind Adscalr.
See the product →