Why not just use AppsFlyer?

If you are already on it and it works, stay. The difference that matters is where ROAS lives: on the large platforms the cost integrations that make ROAS complete sit in the enterprise tier, behind a sales call. Here they are in the free plan, because there is only one product.

Does Traqio measure websites, or only apps?

Both. A product can be an app, a website, or both. On a website the ad click's own id travels to the sign-up, so the attribution is deterministic, as it is on Android. When your app and your website report the same user id, Traqio sees one person.

Do I have to integrate another SDK?

Yes. In an app: one package, an init() call and a track() call. On a website: a small script in the page and one call from your backend at sign-up. We would rather say that plainly than bury it. Six of the Traqio SDKs install from npm today; the others come from a build we send you until they are on their public registries.

Can I trust your numbers if you have no customers?

Do not trust them — check them. Every figure on this page shows its inputs, the attribution is deterministic by default — a probabilistic match runs only if you turn it on, and is labelled as one — and the places our number can drift are listed above in plain language.

2,000 conversions a month free for each product — never pooled across your account — then $0.04 each; an app install and a website's new user count the same. A negotiated rate at high volume. No minimum and no annual contract. No payment provider is wired yet, so nothing is billed during early access.

Other plans are free at my volume. Why pay for this one?

Some free tiers in this category are larger than ours, and our comparison pages say so in each vendor's own words, with the date the page was read. What the price buys here is a different thing: one product on every plan, so nothing is taken away as you grow; a rate published before you talk to anyone; and a number that shows the inputs it was built from.

The MMP that shows its work

Meta says 231.TikTok says 214.Stores say 426.

Three dashboards, three answers, and a spreadsheet where you try to reconcile them. Traqio ties each install and each website sign-up to the ad that paid for it, joins your real revenue to your real spend, and shows you the working.

Get your SDK key See how the number is built

Free up to 2,000 conversions a month. No card, no call.

Ask an assistant about Traqio:ChatGPT Claude Perplexity

Deterministic by default on Android and the web · SKAN and Apple Ads on iOS · Meta CAPI + TikTok Events built in

One day of installs, three dashboards

Example data

Meta Ads Manager

installs claimed

231

TikTok Ads Manager

installs claimed

214

Google Play + App Store

new installs, organic included

426

Meta and TikTok claim 445 installs between them — more than the 426 the stores counted in all.

Traqio

426 installs, each one counted once

Meta118
TikTok151
Apple Ads53
Organic104

Each install is credited to the click it carried, or to organic when no click came with it — then joined to its revenue and to the spend that bought it.

The calculation

ROAS is one division. Here are both of its halves.

Attributed revenue over ad spend, per campaign, ad set and ad. Nothing is modelled and nothing is smoothed — and every figure below adds up, which you are welcome to check.

D30 ROAS · one month of installs

Example data

0.92×

$14,278revenue by D30$15,507ad spend

Attributed installs

9,666

Blended CPI

$1.60

Revenue comes from your app's events, your website's backend and your RevenueCat webhook, joined to the install or the sign-up that earned it. Spend is pulled daily from each ad network's API, in the ad account's own billing currency.

Per campaign

revenue by D30ad spend

Meta·Advantage+ app · US

ROAS 1.34×

2,694 installs · CPI $2.16

$7,788 revenue by D30

$5,812 ad spend

TikTok·Spark Ads · creators

ROAS 1.06×

1,467 installs · CPI $1.50

$2,329 revenue by D30

$2,198 ad spend

Apple Ads·Category keywords

ROAS 0.84×

1,583 installs · CPI $1.66

$2,201 revenue by D30

$2,634 ad spend

Meta·Lookalike 1% · UK, CA, AU

ROAS 0.62×

851 installs · CPI $1.74

$913 revenue by D30

$1,477 ad spend

TikTok·Smart+ · broad

ROAS 0.31×

3,071 installs · CPI $1.10Cheapest installs

$1,047 revenue by D30

$3,386 ad spend

The cheapest installs of the month ($1.10 each) returned 0.31×. Pause that one campaign and the same month reads 1.09× instead of 0.92×.

Spend is stored at all three levels, and every read filters by level. An ad's spend is already inside its ad set's, which is already inside its campaign's — summing the table would triple the denominator and cut every ROAS to a third. There is a regression test for exactly that.

How it works

One install, four hops, deterministic by default.

Our click id is written into your tracking link and read back after the store round-trip. For a click, that is the entire mechanism: no device graph, no cross-app identifier. A probabilistic match happens only if you turn it on, for users who consented in your app — and every install it attributes is labelled probabilistic.

1
The click
Someone taps your ad on Meta or TikTok. Your tracking link carries our click id.
clk_8f2a3c
2
Google Play
The Install Referrer carries that id across the store round-trip — the one place Android hands it back intact.
clk_8f2a3c
3
First open
The SDK reads the referrer once and sends the install — never the pasteboard, and no advertising id unless you switch it on.
clk_8f2a3c
4
The join
The install is matched to campaign, ad set and ad on the network's own ids. Deterministic: same input, same answer, every time.
clk_8f2a3c

iOS works differently and we will not pretend otherwise. Apple limits everyone to SKAdNetwork's aggregated postbacks there, except for installs from its own ads, which Apple attributes one by one — we verify Apple's signature, decode the conversion value against your mapping, and report modelled revenue separately from measured revenue.

On a website

Same click id, and no store in the way.

A tracking link can land on your site. Our click id rides in the landing URL, waits on your own domain and reaches your backend with the sign-up. That is a shorter trip than an app install's, not a less certain one.

1
The click
Someone taps your ad. Your tracking link opens your page, with our click id in the URL.
clk_8f2a3c
2
Your page
A small script keeps that id first-party, on your own domain. No third-party cookie, no fingerprint.
clk_8f2a3c
3
The sign-up
Your backend creates the account and reports it — the user's id and the click id, signed with a secret key.
clk_8f2a3c
4
The join
The new user is tied to campaign, ad set and ad. Their purchases follow them there, whatever page they pay on.
clk_8f2a3c

When the chain breaks — an ad opened in an app's own browser, then a sign-up in another one — the sign-up arrives without its click id and is counted as organic. It is never handed to the nearest campaign.

Why you can believe the number

We would rather show the mechanism than ask for your trust.

We are new, so there are no customer logos to put here and we are not going to borrow anyone else's. What we can do instead is open the box: here is how the number is made, and here is where it can be wrong.

Two senders, one conversion

During a migration you will run us and the network's own SDK at the same time, so every purchase is sent twice. If the two senders disagree on the event id, your conversions double and every ROAS you read is half the truth.

Your app's Meta SDK

sends the purchase from the device

event_id tq1_9c4e17b0a2f8

Traqio's server

sends the same purchase via the Conversions API

event_id tq1_9c4e17b0a2f8

Where our number can be wrong

Every measurement tool has a list like this. Most do not print it on the homepage.

Ad revenue is an estimate: AdMob reports revenue per impression on the device and tells us how precise each figure is. It will drift a few percent from your actual payout. We chose it anyway, because the aggregate Reporting API cannot be attributed to a campaign at all.
Spend moves for about three days: Ad networks restate spend after the fact — late-billed impressions, invalid-traffic credits. We re-pull a rolling window instead of freezing the first number they gave us, so yesterday's ROAS can still shift slightly.
iOS is aggregated, for everyone: SKAdNetwork is coarse, delayed and privacy-thresholded. Anyone promising you device-level iOS attribution is describing something Apple does not allow.
Organic is what is left over: An install we cannot tie to a click is recorded as organic, not assigned to the nearest campaign. That makes organic a real bucket rather than a rounding error, and it is why our attributed totals are conservative.
On the web, the click id lives in one browser: A visitor who taps your ad inside an app's own browser and signs up later in Safari arrives without it, so that sign-up is organic. Safari can also clear an id kept by script after seven days. We count those sign-ups as organic rather than guess where they came from.

Payback

See which campaign pays back, and which one stays in the red.

Cumulative ROAS by days since install, per campaign. A line that crosses 1.0× has earned back its spend. Day 0, the install day, is counted — vendors differ on that; we count it.

Cumulative ROAS per campaign

One month of installs, by days since install

Example data

Campaign	D7	D30	Paid back
Meta · Advantage+ app · US	0.98×	1.34×	D14
TikTok · Spark Ads · creators	0.75×	1.06×	D30
Apple Ads · Category keywords	0.59×	0.84×	not yet
Meta · Lookalike 1% · UK, CA, AU	0.43×	0.62×	not yet
TikTok · Smart+ · broad	0.20×	0.31×	not yet

23% of the month's revenue lands on install day. A cohort table that opens at D1 never shows it to you — it shows a smaller number and calls it the total.

Pricing

Three plans. One product.

The free plan is not a trial and not a preview of the real thing. It is the real thing, with a volume cap. The plans differ in how much you send, never in what you get.

Free

2,000 conversions / mo, per product

Deterministic attribution by default, app installs and website sign-ups
ROAS per campaign, ad set and ad
Meta CAPI + TikTok Events forwarding
SKAN conversion values and postbacks, Apple Ads attribution
RevenueCat and AdMob revenue
Retention and ARPU cohorts from D0

Pay-as-you-go

$0.04 / conversion

after each product's first 2,000, each month

Deterministic attribution by default, app installs and website sign-ups
ROAS per campaign, ad set and ad
Meta CAPI + TikTok Events forwarding
SKAN conversion values and postbacks, Apple Ads attribution
RevenueCat and AdMob revenue
Retention and ARPU cohorts from D0

Enterprise

Custom

volume rate, same product

Deterministic attribution by default, app installs and website sign-ups
ROAS per campaign, ad set and ad
Meta CAPI + TikTok Events forwarding
SKAN conversion values and postbacks, Apple Ads attribution
RevenueCat and AdMob revenue
Retention and ARPU cohorts from D0

Talk to us

That list is identical under all three plans, and that is the whole point. What a larger customer negotiates is the rate — never the product.

One install on mobile, one new user on the web: the same counter, the same 2,000 free a month, for each product.

A conversion is one attributed install in an app, and one new user — a user id seen for the first time — on a website. Events are measured and reported, never billed. Organic installs we simply observe are free. Billing is not switched on during early access, so nothing is charged today.

Straight answers

What Traqio is not yet

We just told you where our number can be wrong, so it would be strange to hide what the product does not do. Here is the honest list.

Android and the web are where attribution is deterministic. On iOS you get SKAN, like everyone else, and Apple Ads' own attribution for installs from Apple's ads — plus, only if you turn it on and your users consent, a probabilistic web-to-app match, labelled as such.
Spend comes from Meta, TikTok and Apple Ads, and from the main game ad networks such as AppLovin. The Google Ads connector is built but waits on Google's approval of our API access, so there is no Google Ads ROAS yet; other networks are not connected.
Our JavaScript app SDKs and the website's browser and server SDKs for JavaScript are on npm; the native SDKs are not on their public registries yet. One app SDK is verified end to end on a real device, and the server SDKs end to end against our own API — the install guide says which is which, and how far each one is proven.
We are not a certified measurement partner at Meta or TikTok yet, so we are not in the MMP drop-down inside your ads manager.
Billing is not wired. The free tier is genuinely free right now, because there is nothing to charge you with.

No dates attached, deliberately. A roadmap with missed dates is worth less than a clear list of what is true today.

FAQ

The questions you were going to ask

Why not just use AppsFlyer?: If you are already on it and it works, stay. The difference that matters is where ROAS lives: on the large platforms the cost integrations that make ROAS complete sit in the enterprise tier, behind a sales call. Here they are in the free plan, because there is only one product.
Does Traqio measure websites, or only apps?: Both. A product can be an app, a website, or both. On a website the ad click's own id travels to the sign-up, so the attribution is deterministic, as it is on Android. When your app and your website report the same user id, Traqio sees one person.
Do I have to integrate another SDK?: Yes. In an app: one package, an init() call and a track() call. On a website: a small script in the page and one call from your backend at sign-up. We would rather say that plainly than bury it. Six of the Traqio SDKs install from npm today; the others come from a build we send you until they are on their public registries.
Can I trust your numbers if you have no customers?: Do not trust them — check them. Every figure on this page shows its inputs, the attribution is deterministic by default — a probabilistic match runs only if you turn it on, and is labelled as one — and the places our number can drift are listed above in plain language.
What does it cost?: 2,000 conversions a month free for each product — never pooled across your account — then $0.04 each; an app install and a website's new user count the same. A negotiated rate at high volume. No minimum and no annual contract. No payment provider is wired yet, so nothing is billed during early access.
Other plans are free at my volume. Why pay for this one?: Some free tiers in this category are larger than ours, and our comparison pages say so in each vendor's own words, with the date the page was read. What the price buys here is a different thing: one product on every plan, so nothing is taken away as you grow; a rate published before you talk to anyone; and a number that shows the inputs it was built from.

Read further

One page per measurement problem

From attributing an install to the campaign that caused it, to tying a website sign-up to its ad click — each use case on its own page.

All use cases How Traqio compares with established MMPs

Point your spend at the ads that actually pay.

Create an account, declare your product — an app, a website or both —, and get your keys. It takes about a minute, and nobody will call you.

Get your SDK key Ask for early access

Free up to 2,000 conversions a month. No card required.

Built and shipped by Khufu FZCO. A real person reads hello@traqio.app.

Meta says 231.TikTok says 214.Stores say 426.

ROAS is one division. Here are both of its halves.

Per campaign

One install, four hops, deterministic by default.

The click

Google Play

First open

The join

Same click id, and no store in the way.

The click

Your page

The sign-up

The join

We would rather show the mechanism than ask for your trust.

Two senders, one conversion

Where our number can be wrong

See which campaign pays back, and which one stays in the red.

Cumulative ROAS per campaign

Three plans. One product.

What Traqio is not yet

The questions you were going to ask

One page per measurement problem

Point your spend at the ads that actually pay.