FTD Magnet AI is live in Push. Every campaign, every advertiser, effective today.
The optimization layer that reshaped how we scale Pop traffic is now available across every Push campaign on the Traffic Nomads Ad Network. Same algorithm. Same deposit-based logic. A completely different format to test it in.
When FTD Magnet AI launched on Pop, the objective was to pull campaign optimization away from click-based signals and closer to what actually generates revenue in iGaming: the first deposit. The results, published in our earlier case study, made the direction obvious.
Today, that same optimization logic goes live in Push. Every advertiser running Push campaigns on Traffic Nomads can now switch on FTD Magnet AI and let the algorithm optimize toward depositor behavior instead of open rates or clicks.
This article covers what we learned in Pop, why Push is the natural next step, what changes on the advertiser side, and how to get started.
iGB Affiliate Awards 2026
FTD Magnet AI was nominated for Best Tech at the iGB Affiliate Awards 2026. The launch on Push is the next step for the same technology now being recognized by the industry.
What FTD Magnet AI Delivered in Pop
Before we get into what changes for Push, a short recap of the numbers our earlier case study produced. All figures come from campaigns run inside the Traffic Nomads Ad Network, comparing FTD Magnet AI optimization against manual bidding on similar offers, GEOs, and traffic conditions.
| Metric | FTD Magnet AI vs Manual Bidding | What It Means |
|---|---|---|
| Share of high-value post-click events (registration → FTD) | 66% generated by AI-optimized campaigns | Most of the funnel progression that leads to monetization |
| Conversion rate | Up to 8x higher | Traffic mix aligns with deposit-ready users |
| Cost per event (registration or FTD) | Up to 5x lower | Budget shifts away from low-intent activity |
The pattern that emerged was consistent. Once the algorithm had enough deposit signal to work with, it stopped chasing cheap conversions and started prioritizing traffic segments that historically produced FTDs. The efficiency gain wasn't marginal — it was structural.
"When optimization follows depositor behavior, the traffic profile starts changing. The system gradually filters out segments that only generate activity and focuses on users that are statistically closer to depositing." — Mateo Brizi, CTO, Traffic Nomads
That result opened an obvious question inside the team: how far does this logic travel? Pop is a specific format with specific funnel dynamics. What about the channels that operate on completely different mechanics?
Why Push Was the Right Next Format
Push is fundamentally different from Pop. Push impressions are delivered to users who have already opted in to receive notifications from a publisher. That single fact changes the optimization problem.
Pop traffic asks the algorithm to identify quality inside broad, cold impression volume. Push asks a sharper question: within a warm, opted-in audience, which subscriber-format-timing combinations are most likely to convert into deposits?
In practice, this is where a lot of Push campaigns leak performance. Media buyers optimize toward open rates or click-through rates because those are the metrics the format naturally surfaces. Both are useful, neither is deposit intent. A subscriber who opens every notification is not necessarily a subscriber who will deposit, and the cost of finding that out through manual iteration is high.
FTD Magnet AI attacks the same problem from the deposit side, using signals specific to how Push behaves.
What FTD Magnet AI Evaluates in Push
The core logic — optimize toward FTDs, not clicks — is identical to what runs in Pop. The signal set is different.
- Subscriber source patterns. Which publisher subscriber lists have historically produced depositors, not just openers.
- Time-of-delivery patterns. Deposit intent inside Push is heavily time-sensitive. The algorithm learns which hours per GEO correlate with FTDs, not just with opens.
- Device and OS combinations. Push conversion behavior varies significantly between iOS and Android web-push, and across mobile browsers. The algorithm treats these as distinct segments.
- Frequency response. Some subscribers respond better to fewer, higher-intent notifications; others are frequency-tolerant. The algorithm balances exposure per subscriber against deposit outcome.
- Landing page interaction depth. A click that bounces in under two seconds is a very different signal from a click that starts the registration flow. FTD Magnet AI weights them accordingly.
All of this happens automatically. Advertisers don't need to build the segmentation logic themselves; they need to make sure the deposit postback is firing correctly so the algorithm has something to learn from.
"The mechanics of Push are different, but the principle is the same. If your optimization stops at the click, you're leaving the most valuable part of the funnel untouched." — Mateo Brizi, CTO, Traffic Nomads
What Changes for Advertisers
The setup is intentionally close to how our Pop advertisers already run FTD Magnet AI. The switch is a single toggle inside the campaign builder, but there are three things that matter more than usual on the advertiser side to make the algorithm work faster.
1. Deposit postback has to be live before the campaign launches
This is the single most important requirement. FTD Magnet AI in Push cannot optimize toward something it can't see. Send the FTD postback from your tracker into Traffic Nomads before the first impression fires. Test with a small verification campaign so you can confirm the deposit event is landing correctly, tied to the right source ID, GEO, and device.
2. Let the algorithm accumulate deposit signal before drawing conclusions
FTD Magnet AI needs deposit data to learn. On Push, we've seen learning stabilize around 20–30 confirmed FTDs per major GEO or offer segment. Running a campaign for two days on a low budget and killing it because the CPA looks off will produce misleading conclusions. The algorithm is still in cold-start.
3. Keep landing pages honest about deposit intent
Because Push subscribers are warmer than Pop impressions, the landing page has to close the loop cleanly. Bonus offers, deposit CTAs, and value propositions need to be visible above the fold. Any friction between the click and the deposit slows down learning for the algorithm and inflates cost per FTD in the interim.
Who Should Test First
FTD Magnet AI in Push will produce the fastest signal for advertisers running consistent Push volume across multiple GEOs, promoting offers with a clean, tracked deposit flow. Sportsbook and casino operators with well-defined welcome bonuses, and affiliates who own tracked funnels through to deposit, will hit the learning threshold quickest.
If Push has historically been a reactivation channel for you and the KPIs stopped at open rate and CTR, the shift will be more visible. Deposit-based optimization tends to expose which subscriber sources were actually contributing to revenue and which were producing noise.
Final Thoughts
Every optimization system that stops at the first measurable event is leaving the most important part of the funnel untouched. In Pop, moving optimization past the click was the change that produced 8x conversion rates and 5x lower cost per event.
Push has always been a strong reactivation and short-funnel format for iGaming. What was missing was a way to align its optimization with the same KPI operators care about — the first deposit. That's what FTD Magnet AI is now doing across every Push campaign on the Traffic Nomads Ad Network.
"We built FTD Magnet AI because dashboards full of clicks and registrations don't pay the bills. Deposits do. Push was the next logical channel because the audience is already opted in — what's left is making sure we're finding the ones who convert." — Mateo Brizi, CTO, Traffic Nomads
The results in Pop set the direction. Push is where advertisers get to generate their own numbers on the same optimization logic. The switch is live now inside every Push campaign on Traffic Nomads.
