It assumes
That without the campaign, both groups would have moved the same way. Matching makes that likelier; it can't make it certain.
Did your campaign change minds? Measure it, honestly
Onflow Brand Lift, built into Onflow Pulse, asks the same question before and after a Paid Promotions campaign, in the channels that ran it and in matched channels that didn't. Comparing the two changes is the fairest reading Telegram allows. The answer may be a lift, no lift, or "not enough to tell", and the report says which.
Already a member? Open Brand Lift
Matched channels for comparison Smallest detectable change shown at the quote Method and limits in plain words
A study, step by step
Two groups of channels, two waves of polls
A study is attached to one of your Paid Promotions campaigns whose bookings are already paid into escrow.
- You
You write the questions
One to three questions, each with the answer you hope to move. You mark one as the main question; the others are read more strictly.
- Onflow
We match control channels
Channels like the campaign's — same category and language, similar size and Onflow Verified score — that haven't run your ads and don't share an owner with them.
- Both groups
Before and after
A first wave of polls closes before the campaign's first post. A second wave runs after each channel's posts, at the same times of day in both groups.
- The report
The report
How much the answer you chose moved in the campaign's channels, minus how much it moved in the others, with a range and how sure it is.
Timing matters. The first wave needs time for channels to accept, a full poll window and a margin before the campaign's first post. If the campaign starts too soon, the quote says so and can't be confirmed; if the campaign is moved earlier later, the study is refused or quoted again.
The method, in plain words
Each channel counts once
- Channel by channel. For each channel we take the share choosing your answer after, minus the share before. A channel with many voters can weigh more, but never more than three times a typical one.
- Compared with control. The lift is the average change in the campaign's channels minus the average change in the control channels. That removes what changed everywhere, like the news or the season.
- How sure. The range accounts for having a limited number of channels, and a shuffle test checks it. Extra questions are held to a stricter bar so asking more doesn't find more by chance.
- Minimums. At least 8 channels in each group and 200 voters in each wave of each group. Below that, the study is inconclusive and the report lists what was missing.
What it can and can't tell you
Assumptions and limits, stated up front
Voters choose to vote
Polls are anonymous and voluntary, and voters are never rewarded. The people who answer may differ from the readers who don't, and the same person may or may not answer both waves.
Small changes can hide
The quote shows the smallest change the study could reliably detect, from past polls (or a cautious default, labelled as such). A real change smaller than that may read as no lift.
What it costs
A study fee, plus the channels' poll fees
The study fee is $49. On top, you authorise a cap for the polls in both groups and both waves, with each channel's fee set by its size band, plus the platform fee. Both are charged when you confirm. Onflow Credit can pay only the study fee and the platform fee; the channels' fees come from the wallet, and a study the wallet can't cover can't start. Brand Lift is included with Elite, Master and Ultimate.
Refunded with the study fee
Everything, if you cancel before the first polls go out. If the campaign is cancelled before its posts run, or a fault on our side stops the study, the study fee and everything not yet used come back, and the polls that already ran are paid. If a fault on our side means the study can't produce its result, the study fee comes back with the fees for the polls that failed.
Refunded per channel
As in Pulse: a channel that didn't accept, or whose poll failed on our side, is refunded with its platform fee. A channel that deleted a poll early, or an upheld dispute, gives back the channel's fee; Onflow keeps the platform fee.
Kept
The study fee, when the study ends inconclusive because too few channels or voters took part. The quote says so before you confirm.
| Plan | Brand Lift | Off the study fee |
|---|---|---|
| Free | No | 0% |
| Lite | No | 3% |
| Plus | No | 6% |
| Elite | Yes | 10% |
| Master | Yes | 14% |
| Ultimate | Yes | 18% |
The plan discount comes off the study fee only, never off the channels' poll fees, the platform fee or any payout, and it is never added to a coupon: the better of the two applies. A plan never changes a study's result.
Paying the channels. Each wave's host is paid as in Pulse: 7 days after its poll closes, unless you dispute that payout within 72 hours; accounts in a lower reliability band wait longer. You may release a live poll's payout early on the website, but that order then has no dispute and no refund, including the Brand Lift wave refund. How payouts work.
No promised result. A study measures your campaign; it can't guarantee that it worked. Want a quick read on one question instead? Try Onflow Pulse.
Inside your dashboard
The quote before you pay, the report after
View full picture
View full report
Before you start
Common questions
Will a study show that my campaign worked?
Not necessarily. A study measures; it does not promise anything. It can show a lift, no detectable lift, or a drop, and it can end inconclusive if too few channels or voters took part.
What does inconclusive mean, and do I get the fee back?
Inconclusive means a group ended with fewer channels or fewer voters than the minimum, so no fair comparison could be made. The report says exactly what was missing. The study fee is kept in that case, and the quote says so before you confirm.
When is the study fee refunded?
With everything else, if you cancel before the first polls go out. If the campaign is cancelled before its posts run, or a fault on our side stops the study, the study fee and everything not yet used come back, but the polls that already ran are paid. It also comes back if a fault on our side means the study can't produce its result. A channel that deletes a poll early has its fee refunded as in Pulse, and Onflow keeps the platform fee.
When are the channels that host the polls paid?
As in Pulse: 7 days after each poll closes, unless you dispute that channel's payout within 72 hours of the close. Accounts in a lower reliability band wait longer. If you release a wave's payout early, you give up that poll's refund, including the Brand Lift wave refund.
Who sees how people voted?
Nobody. Every poll is an anonymous Telegram poll, and the study uses the totals for each answer in each channel, never who voted.
Which plans include Brand Lift?
Elite, Master and Ultimate. The study is paid for separately: Onflow Credit can pay the study fee and the platform fee, the channels' poll fees come from the wallet, and your plan's discount on Onflow's own services applies to the study fee, never to the channels' poll fees.
Measure your next campaign
Open a Paid Promotions campaign and choose Measure brand lift to see the quote, the timing and the smallest change it could detect, before anything is charged.