How GPT sites make money, from the operator’s side

GPT sites make money from the gap between what offerwall and survey partners pay for a completed offer and the reward the site passes to the member. Partners pay the site per validated conversion. The site credits the member a share in points, pays out when they redeem, and keeps the rest after payout fees and reversals.

Most articles on how GPT sites make money are written for members who want to earn. This one is written for operators. It treats the GPT site business model as a profit and loss statement: revenue, the costs that scale with it, the timing of cash, and what the business is worth at the end.

Every dollar on a get-paid-to site starts with an advertiser who wants an action: an app install, a game played to a certain level, a survey completed, a trial started. The advertiser pays a network, the network pays your site, and your site pays your member.

The member sees 840 points. You see $0.36 that still has to pay for the cash-out fee, any reversed offers, support and software. That gap is small per offer, which is why get-paid-to site revenue depends on volume.

Two splits sit between the advertiser and your member, and people often confuse them. The first is between the network and you, and the network sets it. The second is between you and your member, and you set it.

You cannot change the first split much, beyond choosing partners that pay well for your audience. The second is your main lever. A higher member share makes the site more attractive, and members do compare: the same offer often pays different amounts on different GPT sites because each site sets its own margin. A lower share keeps more per offer but can cost you the members who generate volume.

Setting the share per partner, rather than once for the whole site, lets you compete where members compare most and keep more where they do not.

A bonus on top of the normal share, for one partner and for a limited time, is a cheaper way to move members than raising the share across the board. It can push activity toward a partner that pays well for your audience, or help a new partner build enough volume to judge it. Because the bonus is a cost on every conversion it touches, give it an end date and compare the margin before and after.

Offer margin is the core, but it is not the only line. Each of the others comes with conditions.

These are the costs of running a GPT site, in rough order of size.

The largest cost by far, and the one you set yourself through the member share. Everything else on this list is paid out of what remains.

Networks validate conversions after the fact. RapidoReach’s 2020 guide describes networks taking 30, 60 or 90 days to decide whether to pay. When a conversion is reversed after the member has cashed out, the site absorbs the loss, because the reward has already been paid. RevBoost suggests investigating any network whose reversal rate goes above 5 to 8%. Holds on new earnings and manual review of first withdrawals are how operators keep this line small.

Support time goes mostly to missing-credit questions: a member completed an offer and the points did not arrive. Software is a script and its hosting, a custom build, or a platform licence. Marketing is optional in principle and usually the line that decides growth.

A GPT site can be profitable on paper and still run out of cash. Members expect withdrawals within days, while many networks pay monthly. Monlix, for example, states in its FAQ that it pays NET30, once a month around the 15th, with a $20 minimum.

In practice you pay members out of your own funds for several weeks before partner money arrives. A site that grows quickly needs more of this float, not less. Plan for it before you spend on growth, and keep a reserve for conversions that are reversed after you have paid.

Minimum cash-out amounts and holds are cash flow tools as much as fraud tools. A higher minimum means fewer, larger payouts, so fixed fees take a smaller share of each one. A hold on new earnings keeps money in your account until partners have had time to report reversals. Both should be stated in your terms, so members know them before they ask to withdraw.

Referral programs are the most common growth tool on GPT sites. A member shares a link, and when the referred member earns, the referrer receives a commission on top. That commission comes out of your margin; the referred member’s reward is not reduced to pay for it.

Market examples show two models. Swagbucks is described as paying 10% of a referred member’s earnings for as long as they stay active, by Savvy New Canadians, a secondary source. Freecash now pays fixed rewards instead: its help centre lists $2.50 in the US when a referred member signs up and installs a first game, and $10 at their first withdrawal. A percentage grows with what the referred member earns; a fixed reward costs the same whatever they earn later.

A referral is worth paying for when the referred member’s offers leave enough margin after the commission. For example, on a 30% margin, a 10% commission on what the referred member earns is affordable. On a thin margin with high reversals, it may not be.

A monthly statement shows the economics most clearly. The table below is an illustration, not a forecast. The numbers are round and chosen to show how the lines relate. Your own will depend on your audience, partners and countries.

Two things stand out. First, the member share dominates: moving it from 65% to 75% removes $1,000, more than half the profit. Second, the fixed costs are small next to the variable ones, so the decisions that matter most are the share, the hold rules and the choice of partners.

Small GPT sites change hands on marketplaces. One Flippa listing for a long-running incentive site asked $10,800, with seller-reported figures of about $2,000 a month in revenue and $900 a month in profit from more than 10,000 members. That asking price is roughly 12 times monthly profit. Seller figures are unverified, and an asking price is not a sale price.

At the large end, Gain.gg reported more than 375,000 active users and over $1.75 million paid out to members in a Mistplay case study. Freecash reported over 80 million users and more than $300 million paid out, according to PocketGamer.biz. Paid out is not profit, but it shows how much volume the model can carry.

If you are buying a site, or building one to sell later, these are the figures a buyer will ask to see:

Software priced as a share of revenue costs more each month as partner revenue grows, whether or not your margin grows with it. A fixed price does not.

RewardsLabel charges $149 a month or $1,490 a year, plus a $25 activation fee, with no revenue share. In the sample P&L above, software stays at $149 whether partner revenue is $10,000 or $100,000. Every other line remains yours: you set the member share in each partner’s currency settings and record it in the console, approve each cash-out from your redemption queue, and pay members from your own accounts. See pricing, or read how to start a GPT site for the launch steps.