When a Contest’s Follower Spike Isn’t the Real Result
A case study opens with a follower count that nearly doubled in two weeks, and the mechanism it credits is a giveaway. What decides whether it means anything is not whether the spike happened, but whether it was still there once the prize was gone, and that is usually the measurement omitted.
The word ‘grew’ is hiding two different claims
A follower count is a stock: a single running total, readable in one glance. Engagement rate, reach and repeat purchase are flows, rates of behaviour measured across a period. A tactic can move one without touching the other, and neither outcome contradicts the other. A contest is engineered to move a stock, and adding names to a running total is exactly the job it is built for. But a flow is a rate, and adding people to its denominator does not raise it. It can lower it. A case study saying an account “grew” makes a stock claim and invites you to hear a flow claim.
What a giveaway actually purchases
An entrant performs a specified action in exchange for a chance at a prize. Commonly that action is following the account, often it also includes tagging other people.
Consider what that selects for. An organic follower arrived because the account was worth seeing again, a weak filter but one pointed at the brand. A contest entrant was screened for wanting the prize. This is why reversion is expected rather than surprising: the audience was recruited against a criterion that expires when the prize is drawn.
One distinction matters here. A real person who followed to enter a draw and unfollowed a month later is not a bot and not part of a coordinated scheme. That is a separate problem. Every entrant can be genuine and the number still reverts.
The question almost no case study answers
The missing data point is specific and small: a second measurement of the same metric, taken after the incentive ended, at a stated interval.
That absence is not neutral. A case study confident the gain had held would have an easy time reporting it: the account still exists and the number is on the profile. A strong retention figure sells better than a peak. When something that cheap is systematically absent, the absence carries information.
This differs from the question about baselines. A baseline asks what the number was before the campaign, a point to the left on the timeline. This asks what it was after the incentive stopped, a point to the right. A case study can disclose an honest starting figure and still say nothing about durability, because the number no case study can show you is often the one taken last.
A hypothetical case, worked through three points in time
The numbers below are invented for illustration, not a reported result from any real account or agency. Picture a hypothetical consumer goods account at 18,400 followers. It runs a fourteen day giveaway, and on the closing day the count reads 41,200.
| Point in time | Follower count (hypothetical) | Gain over baseline |
|---|---|---|
| Day 0, before the contest | 18,400 | baseline |
| Day 14, contest close | 41,200 | 22,800 |
| Day 90, 76 days after the prize | 24,600 | 6,200 |
Of the 22,800 gained over baseline, 6,200 were still present at day 90: 27 percent of the gain retained, 73 percent reverted. The account is genuinely larger, but the headline figure of 22,800 describes a state it occupied briefly. These remain hypothetical numbers, the shape of a reverting metric, not evidence about any real contest.
Three questions to ask before crediting the tactic
- Was the metric measured again after the incentive ended, and is the interval disclosed? Look for a stated date or number of days, not a vague “sustained growth.”
- Is the incentive-acquired cohort reported separately, or blended into one number? A case study giving the unfollow behaviour of the contest cohort specifically has done real analysis. One reporting total account followers has averaged that cohort in with everyone else.
- Did any secondary metric move with the count? Engagement rate, reach, or a business outcome moving alongside the follower total is evidence that something beyond the stock changed. If only the count moved, only the count moved.
None of this makes the tactic illegitimate
A fast, incentive-driven spike is a reasonable thing to want and to report. The problem is never the tactic, it is the distance between what happened and the claim made about it. A case study disclosing a re-check date and an honest retention figure is more credible than one reporting only the peak, whatever that later figure turns out to be. A document that tells you what share of the gain was still there at a stated later date earns trust that a peak followed by silence does not. It is the same move as recognising that one viral post is not the same claim as a strong campaign.
Read a few side by side
These patterns get easier to spot with practice. Browse the case study library, or submit your own. Arriving fresh, how to read a marketing case study without being misled is the framework behind this one.
FAQ
Does a follower count dropping after a contest mean the contest failed?
Not necessarily. Judge it against what the contest was stated to achieve. A temporary expansion of reach is a different goal from durable community growth, and a partial reversion is the expected shape of the first, not evidence of failure at something the tactic never claimed.
How long after a contest should a case study re-check the number?
There is no correct interval to demand, and any number of days asserted here would be invented. Notice instead whether an interval is disclosed at all. A case study that names when it looked again has made its claim testable. One reporting a peak and no second date has not, and that absence is the signal, not the length of the gap.
Are giveaway participants fake accounts or bots?
Usually not, and the two are worth separating. A genuine person who followed for a prize and later unfollowed is a different phenomenon from a fake account created to inflate a count, which is what platform inauthentic-behaviour policy targets. The argument here holds even when every entrant was real and simply not interested in the brand.
Sources
- The number no case study can show you, fetched 2026-09-06.
- One viral post is not the same claim as a strong campaign, fetched 2026-09-06.
- How to read a marketing case study without being misled, fetched 2026-09-06.
