In arbitrage, you live and die by your data. If you cannot see which ad, which placement, and which audience made the money, you are just gambling with extra steps.
This article covers what to actually track and why. Get this right and optimizing becomes obvious. You cut what loses, scale what wins, and stop guessing.
Why do I need a tracker at all?
The ad platform tells you what it spent. The offer network tells you what converted. But neither one connects the two at the level you need. A tracker sits in the middle and ties every conversion back to the exact click that made it.
That connection is everything. Without it you might know your campaign made money overall, but not that one placement is carrying three placements that lose. You cannot fix what you cannot see.
A tracker also catches bad traffic. If one source sends tons of clicks and zero conversions, that is your signal to cut it. The tracker is what turns arbitrage from luck into a process you can repeat.
What data points actually matter?
Do not drown in columns. These are the ones I watch.
- Spend: what you paid for traffic, broken down by source.
- Clicks and CTR: are people clicking, and how many.
- Conversions and CVR: how many clicks turned into the paid action.
- Cost per conversion: spend divided by conversions. This is your core cost number.
- Revenue and ROI: what the offer paid you versus what you spent. ROI is the number that decides if you keep going.
The trick is to look at all of these per source, not just for the whole campaign. A blended ROI of break-even can hide one source making 50 percent and another losing 50 percent. Cut the loser, scale the winner, and suddenly you are profitable.
How granular should my tracking be?
As granular as you can act on. The whole point is to find the winners and losers inside the campaign.
I want to break performance down by:
- Creative or ad: which specific ad drove the conversion.
- Placement or site: which spot or website the traffic came from.
- Audience or targeting: which segment converted.
- Time and geo when it matters: some offers convert better in certain countries or hours.
Use tracking tokens so every click carries its source info into the tracker. Most ad platforms let you pass these automatically. Set them up once at the start. Going back to add them later is painful, and the data you missed is gone for good.
One warning: do not get so granular that every slice has three clicks in it. You still need enough volume per slice to trust the numbers. Group small sources together until they earn the right to their own view.
What tracking mistakes cost beginners money?
The first is running blind for the first few days because the tracker was not set up before launch. Set it up first, test that conversions fire, then turn on traffic.
The second is trusting blended numbers. If you only look at the campaign total, you keep funding losers hidden inside a winning average. Always break it down.
The third is acting on tiny samples. One conversion on a new source does not make it a winner. Wait for enough data before you scale, or you will scale up noise.
The last is not checking that tracking is accurate. Fire a test conversion and confirm it shows up correctly. If your tracker miscounts, every decision you make on top of it is wrong. Trust your data only after you have verified it works.
Key takeaways
- Set up your tracker and verify it before you spend a dollar on traffic.
- Track spend, conversions, cost per conversion, and ROI per source, not just the blended total.
- Break data down by ad, placement, and audience, but keep enough volume per slice to trust it.