What does Meta require for a Conversion Lift study on B2B lead ads, and is my budget big enough?

TofuAnswer by the Tofu team · checked

Meta recommends at least $5,000 and 28 days for a Conversion Lift test, which you create in Experiments in Ads Manager if your account qualifies, with web events sent through the Conversions API. Two 2026 guides citing Meta add a campaign that started in the past year with $5,000 spent and 500 conversions. For B2B lead ads, the conversion count is often the harder bar: at $100 a lead, $5,000 buys 50 leads. If the Get started button is greyed out, you likely haven't met the spend or conversion bar.

Ask Tofu next

Pick a question and Tofu Ads Agent answers it in chat.

As of November 2026.

Can I run a lift study myself, or do I need a Meta rep?

You can run it yourself if your account qualifies. Meta says, "If eligible, you can run conversion lift tests, brand lift tests, or A/B testing in the Meta Experiments tool." In Ads Manager, open All tools, select Experiments under Analyze and report, click Create test, then click Get started under Conversion Lift. The API route is narrower: Meta's developer guide says Conversion Lift Measurement "is currently limited" and tells you to contact your Meta representative.

What does Meta require before a test can start?

Conversions API data, enough spend and conversions, and a setup you won't change halfway. Here is what Meta's help and developer pages say:

RequirementWhat Meta says
TrackingTests that measure standard web events from October 1, 2021 "require advertisers to implement Conversions API"
Spend and conversionsThe Get started button "may be disabled if you do not meet the ad spend or conversion requirements"
What gets testedAd accounts, campaigns or ad sets. If you pick an ad account, "the test will include all active campaigns" in it
The outcomeUnder Conversion Source, you pick a data source and the events to measure. In the API, objectives can't be changed once the study starts
GroupsMeta splits people into a test group and a holdout group, and they stay in their group "for the entire test duration"
ReachYour ads must reach enough people for a significant result. Meta says some tests show spend minimums or recommended spend while you set them up

Is my budget big enough?

Check two numbers: spend and conversions. Meta's setup page says, "We recommend a minimum budget of $5,000 USD and a duration of at least 28 days." AdAdvisor (October 1, 2026), citing Meta's eligibility guide, says you need "a campaign started in the past year with $5,000 spend and 500 conversions." meta-ads.si (sources checked October 2, 2026) gives the same bar, which it says Meta offers as a guide for self-serve tests. It adds that conversions count under a 1-day click, 7-day click or 1-day view setting, and that the bar is prorated for campaigns over 90 days (a 180-day campaign needs $10,000 and 1,000 conversions).

Meta's setup page links to "About Conversion Lift" for these requirements, but that page showed only its title when we checked, so confirm the numbers in Experiments. For B2B, work back from your cost per lead: at $100 a lead, 500 leads cost about $50,000.

How many leads a week does a lift study need to show a result?

Meta publishes no weekly number. With the proration meta-ads.si describes, the bar is about 500 conversions per 90 days, or roughly 40 a week, and that only gets you into the tool. It doesn't promise a clear result. Meta reports confidence as "The probability that ads in this test cell generate incremental conversions greater than zero," with a 90% credible interval around the incremental conversions. As with any experiment, fewer conversions in the test window mean a wider range and lower confidence. Meta suggests "starting with an account level test that looks at the overall effect of your advertising," which also pools the conversions from all your active campaigns.

Can a lift study use qualified leads from my CRM as its outcome?

Possibly, but Meta's docs don't say so directly. Its lift guide lists what a study can measure: Conversions API-based pixels, app events, offline conversion data sets and custom conversions. It also says "event breakdowns for Offline Conversion are not supported." Meta's CRM integration sends lead stages from your CRM to a Meta Pixel through the Conversions API, so a stage such as "qualified" may appear under Conversion Source. Check that list before you plan around it. If the CRM feed isn't set up yet, see sending HubSpot or Salesforce lead stages to Meta.

CRM stages often happen days after the lead. In the API, a post-test window (observation_end_time) lets Meta "continue to match conversions" to each group after the active period ends. During that window, ads deliver normally to both groups.

Can it measure leads from Instant Forms?

Meta doesn't say. Its setup page calls Conversion Lift "particularly useful for businesses that have a website or app," and its list of measurement sources doesn't name Instant Form leads. For Instant Forms, the CRM route above is the documented way to get lead outcomes into a Meta Pixel. Meta's fit guidelines for its Conversions API for CRM start with using Lead Ads (Instant Forms) and getting at least 200 leads a month. If you send people to a website form, a website Lead event (fb_pixel_lead in the API) is a listed source. For the form choice itself, see Instant Forms vs your website form.

What happens to my leads while the test runs?

Some people who would have seen your ads won't. Meta's holdout group is the people who "match your ad audience but are intentionally withheld from seeing your ads." In the API, a single test group's treatment and control percentages must add up to 100, and once the study starts you can't change the split or remove campaigns from it. So run the test when you can afford some lost reach, and don't rebuild the campaigns mid-test.

How do I read the results?

Look at incremental conversions, confidence and cost per incremental conversion. Meta's lift glossary defines incremental conversions as "The estimated number of incremental conversions that occurred as a result of your ads in this test." Cost per incremental conversion is "the average cost of each additional conversion generated by your test ads." Meta's API guide says breakdowns need "at least 100 conversions from test and control groups combined" to show, so a small B2B test may show only the total. Then compare that cost with what a qualified lead is worth to you.

What if I don't qualify yet?

Use a lighter test while you build volume. Meta's A/B test lets you "compare two versions of an ad strategy," such as images or text, and makes sure "nobody sees both." It shows which version did better, not what the ads caused. For a model-based alternative, our page on incremental attribution for B2B lead gen compares it with a lift study, and custom attribution covers counting conversions from your own system. If Meta can't learn from your few leads at all, see what to optimize for with a few leads a week.

Can I create a lift study through the API or Meta's ads MCP server?

Only with access. Meta's developer guide says API access "is currently limited." Meta's ads MCP server lists a tool to "Create a new conversion lift study on an ad account" and one to get its results. Its eligibility check is described for A/B testing, and Meta doesn't say the MCP tool skips the access limit. See Meta's Ads MCP server to connect it.

Sources (13)

Related answers

More on