Giveaway ROI: The Seven Numbers Worth Tracking

Because everything that looks good on day one is temporary by construction.

Entries, reach, new followers — all three peak within twenty-four hours of the announcement and then decay. Reading them at that moment measures the peak, not the outcome, and the peak is exactly what a giveaway is designed to produce. Of course it looks good.

Thirty days is the honest window because it is long enough for the two decays that matter to have happened: the follower churn from people who entered for the prize, and the reach correction the algorithm applies once those followers stop interacting.

Not followers gained. Followers kept.

Take the count the day before the giveaway launched and the count thirty days after the announcement. The difference is what you actually recruited. The number on announcement day is a different, much larger, and much less useful figure.

A retention rate around a fifth to a quarter of the announcement-day gain is ordinary. Much higher usually means a small, well-targeted prize; much lower means the prize was generic.

The one that decides the verdict.

Take your average interactions-per-post over the month before and the month after, each divided by your follower count. If the after figure is lower, the giveaway recruited people who do not care what you publish — and it will keep costing you, because the algorithm serves your posts to a share of your audience based on how the first viewers respond.

This is the most important signal in the list and the most ignored. A giveaway that grew your audience and lowered this number made you less visible than you were before.

The lever is the prize. A prize anyone would want brings people who want the prize; a prize your customer would want brings people who want what you sell. The prize decision is where this is won or lost, before anything is published.

Not entries. Entries divided by the audience you already had.

The raw entry count depends mostly on how far the post travelled, which depends on the prize and on luck. The ratio tells you something the count cannot: whether the people who already follow you found the prize worth ten seconds.

A low ratio with a high entry count means the post went well outside your audience — sometimes what you wanted, often not.

Fifteen minutes of actually reading them.

This is not a metric, and it is more useful than three of the ones that are. Read a hundred comments and sort them roughly: people answering the question, people writing "done", and giveaway-hunter accounts.

If the middle two dominate, your question was not worth answering. If the last dominates, your prize was too generic. Either way you learn what to change, which a number alone never tells you.

The only figure that connects a giveaway to money without guesswork.

Send everyone who did not win a reason to come back within forty-eight hours — a discount code, early access, an invitation — and count how many use it. Unlike traffic or attribution, this is unambiguous: someone entered your giveaway and then bought something.

It is also the step almost nobody takes, which makes it the cheapest improvement available to most giveaways. Growing followers without losing them covers what to send.

Compared with your normal posts, not with the giveaway post.

The announcement is the most-watched post of the operation, because every entrant is waiting for it. If it underperforms your normal reach, something went wrong upstream: the giveaway ran too long and people forgot, or the announcement came days late.

It is a diagnostic rather than a goal. A weak announcement usually means a scheduling problem you can fix next time for free.

The prize at cost, plus shipping, plus your time.

The prize is cheaper than its retail value if you gave away your own product. Shipping is the line item people forget, and on a small item it routinely exceeds the prize. Your time is real: an hour to prepare, an hour spread across the run, half an hour to draw and announce.

Put that total against the retained followers and the redemptions. Most giveaways look far better on this arithmetic than organisers expect — and the ones that do not are almost always the ones with a generic prize.

A giveaway that brings 400 followers, keeps 90, and drops engagement by a third has failed.

That sentence is the whole article. The 400 is what gets screenshotted and shared internally; the 90 is what you actually bought; and the engagement drop is the bill that arrives later, quietly, in the form of every subsequent post reaching fewer people.

The good news is that the failure mode is singular and preventable: it is nearly always the prize. How to pick a winner nobody disputes handles the credibility side, but no amount of proof rescues a giveaway that recruited the wrong people.

How many entries should I expect? It depends entirely on account size and prize appeal. Be suspicious of promised numbers, and use entries per follower rather than the raw count.

Is a follower drop after the announcement normal? Yes, especially if you required a follow. Measure at thirty days, not at forty-eight hours.

Can I attribute sales to a giveaway? Only loosely, except through a consolation code, which is unambiguous. Use that rather than trying to model attribution.

What engagement drop is acceptable? None is good, but a small dip that recovers within a month is normal. A drop that persists past thirty days means the prize was wrong.

Should I run fewer, bigger giveaways? Usually yes. One a quarter for a brand, at most one a month for a local business — past that the audience learns to react only to giveaways.

What is the single best improvement? The consolation offer to non-winners within forty-eight hours. It costs almost nothing and it is the only step that reliably turns entrants into customers.

Lukky draws from the comments on an Instagram or Facebook post, removes duplicates, and produces a certified result you can publish. See how it works.