The highest payout in the catalogue is rarely the one that earns you the most. This guide breaks down what each offer type asks of your visitor, which traffic each one suits, and how to compare offers using the only number that matters.
Open any offer catalogue and the first thing you notice is the payout column. A $14.00 offer sits three rows above a $0.35 offer, and the conclusion seems obvious.
It is usually wrong. The $0.35 offer is a one-tap install that converts on a quarter of the traffic sent to it. The $14.00 offer requires a credit card. Whichever one earns you more depends entirely on who is clicking, and the payout column cannot tell you that.
This guide explains what each offer type actually asks of a visitor, and how to compare them properly.
CPA, cost per action. The umbrella term, and also the label used for offers where the action is something specific: a registration, a form submission, a subscription, a purchase. The advertiser pays when that named event fires.
CPI, cost per install. The visitor installs an app and opens it. That is the whole requirement. Payouts are usually small, conversion rates are usually high, and mobile traffic converts on it at rates no other type matches.
CPE, cost per engagement. The visitor installs and then does something inside the app: reaches a level, completes a tutorial, plays for a set time, makes a first deposit. Payouts are several times a CPI on the same app, because the advertiser is buying a retained user rather than an install.
CPL, cost per lead. The visitor submits contact details. Often an email address and a zip code, sometimes a full form. Payouts sit between CPI and CPA, and geo restrictions are usually tight.
You will also see offers described as SOI (single opt-in, email only) and DOI (double opt-in, email plus confirmation click). DOI pays more and converts at roughly half the rate. Neither is better; they suit different audiences.
Payout tracks friction almost perfectly. The real question when choosing an offer is how much friction your specific traffic will tolerate.
| Type | What the visitor does | Typical friction | Best-suited traffic |
|---|---|---|---|
| CPI | Install and open an app | Very low | Mobile, gaming, tier 2 and 3 |
| CPE | Install and reach a milestone | Medium | Engaged gaming audiences |
| SOI lead | Enter an email address | Low | Broad, all tiers |
| DOI lead | Enter email, confirm in inbox | Medium | Desktop, older audiences |
| CPA signup | Register an account | Medium to high | Intent-driven, tier 1 |
| CPA purchase | Pay for something | Very high | Buyer intent only |
Read that table against your own audience and most of your offer selection answers itself. A gaming audience on mobile will complete CPI and CPE offers all day and abandon a credit card form. A niche desktop audience that arrived from a comparison article will do the opposite.
EPC, earnings per click. Payout multiplied by conversion rate.
Offer B earns 37% more per click, and it will keep doing so until your traffic mix changes. This is why experienced publishers glance at the payout column and then ignore it.
Two cautions. First, the EPC shown in a catalogue is a network-wide average across every publisher's traffic, not a prediction about yours. It is a starting point for shortlisting, not a forecast. Second, EPC on small samples is noise. A hundred clicks tells you nothing; a few thousand starts to tell you something.
Test payouts against your own traffic, never against the catalogue. The offer that pays a competitor best is a fact about their audience, not about the offer.
Every offer carries constraints, and ignoring them is the most common cause of "my conversions are not tracking".
Countries. A visitor from outside the allowed list will not convert, no matter what they do. If your traffic is international, this alone argues for a SmartLink over a hand-picked offer.
Devices and operating systems. An iOS install offer cannot convert for an Android user. Check the OS list, not just "mobile".
Conversion event. Read what actually triggers payment. An offer described as an install that pays on level 5 is a CPE offer wearing a CPI label, and your conversion rate will be a fraction of what you expected.
Incent flag. This is the one that ends accounts. If an offer is not marked incent-friendly, you cannot reward users for completing it. Running non-incent offers on an offerwall or a rewarded locker produces conversions that get charged back and offers that get pulled from the whole network.
Publishers often assume geo pricing is arbitrary. It is not. An advertiser pays what a user is worth to them, and that value varies enormously by market: purchasing power, average revenue per user, competition for ad inventory, and how expensive that user is to acquire elsewhere.
The practical consequence is that a single app can appear as a $4.20 offer in Germany and a $0.28 offer in Indonesia, and both prices are rational. What matters to you is not the price but the ratio: German traffic converts on that offer at a fraction of the Indonesian rate, and the two EPCs are often closer than the payouts suggest.
This is also why tier 3 traffic is not the poor relation it is assumed to be. Low payouts with high completion rates and enormous volume can out-earn a small quantity of expensive tier 1 clicks, particularly on CPI offers where the friction is a single tap. The publishers who do best with international traffic are the ones who stopped filtering their reports by payout and started sorting by EPC.
There is a case for the expensive offers, and it is narrower than most publishers hope.
Take the high-friction CPA offer when the traffic arrived with matching intent. A visitor who came from a "best VPN for streaming" article and lands on a VPN trial signup is a genuinely different visitor from one who clicked a generic download button. Intent that specific converts at rates that make the payout worth chasing.
Take it also when your locked content is valuable enough to justify a real ask. Nobody fills in a card form for a wallpaper pack. Some people will for a complete course, a paid tool licence or a full archive.
Outside those two cases, the expensive offer is a trap: it takes the same traffic, converts a tenth as often, and pays five times as much. That arithmetic loses.
The delivery surface changes which types perform, sometimes dramatically.
Content lockers favour short, low-friction offers. The visitor is trying to get to something else, and every extra step costs completions. SOI leads and CPI offers dominate here. High-payout CPA offers work only when the locked content is genuinely valuable enough to justify the ask.
Offerwalls favour CPE. The user is inside your economy, they are already engaged with games, and they will happily reach level 10 for a large currency reward. This is the one surface where the highest-payout offers routinely also produce the best EPC.
SmartLinks sort this out automatically, which is precisely the argument for using them when your traffic is mixed or unknown.
That loop takes an hour a month and it is the difference between a publisher whose revenue drifts down and one whose revenue compounds.
If you are earning your first few hundred dollars, the biggest lever is not offer selection. It is traffic volume and placement quality. A perfectly chosen offer on a badly placed locker earns less than a mediocre offer on a great one.
Get the surface right, get the audience right, then tune the offer mix. In that order.
There is a version of this mistake that looks like diligence: rotating offers every two days in search of a better one. All that produces is samples too small to compare, and a publisher who has changed six variables and can attribute the result to none of them. Change one thing, give it enough traffic to mean something, then change the next.
The offer catalogue is visible as soon as you have an account, with live EPC and conversion rate on every campaign, filtered to the geos and devices your traffic actually comes from.
Create your ToroAds account to see the catalogue, or read the platform overview if you are still working out how the pieces fit together.
Continue reading from the ToroAds blog.
