Traffic Test with Native Ads: $3,547 on Diabetes in Mexico #lazycasestudy

The affiliate media space is currently seeing growing demand for alternative traffic sources. For instance, individual affiliate programs, agency sellers, and influencers are actively promoting the topic of native networks. Recently, we published a piece featuring representatives from the AdsKeeper ad network, where we broke down in detail how to run nutra on their traffic. And today we're sharing a case study from one of our media buyers who ran native traffic for the first time — and went straight into profit!

Campaign Overview

• Offer: Diapsula
• Vertical: diabetes
• Affiliate program: OMNI CPA
• Traffic source: native network AdsKeeper
• Campaign period: 12.04.2025 – 12.05.2025
• GEO: Mexico
• Spent: $2,661.66
• Revenue: $3,547.00
• Profit: $885.34
• ROI: 33%

The goal was straightforward — test native traffic in the diabetes niche targeting Mexico. The Diapsula offer from OMNI CPA was used, and the funnel had already been tested on another traffic source, so the main question was how relevant AdsKeeper would be for this type of case. I'll also note that the collaboration experience was genuinely smooth: a manager was assigned right away and provided useful advice throughout the entire campaign, the landing page sailed through pre-moderation with no revision requests, no additional conditions, and no cloaking required. That alone saved a significant amount of time at the start.

The entire setup came down to generating a tracking link in the affiliate program, connecting our own pre-lander, and launching. No headaches with creatives, no "optimization" to fit their moderation policy — overall, they let us enter the traffic source like normal human beings.

Creatives and Pre-landers

As mentioned above, the funnel was already ready, but on the manager's advice we added another pre-lander to the split test. One featured was an interview with a well-known scientist 

Spoiler: the pre-lander featuring the Japanese specialist delivered an average conversion rate 20% higher.

☝️ Various door and medicine cabinet creatives were also tested. In AdsKeeper, the "doors" performed better — unlike another traffic source where medicine cabinets drove a higher CTR.

As for teasers (creatives) — a standard teaser approach appealing to an unexpected cause of the disease. The ad copy used was:

• "Endocrinólogo: 'Bebe esto para eliminar la diabetes en 7 días.'" — translated as: "Endocrinologist: 'Drink this to eliminate diabetes in 7 days.'"
• "¡La diabetes no es por lo dulce! Lee aquí la principal causa" — translated as: "Diabetes isn't caused by sugar! Find out the real cause here." This line was used in teasers featuring a cache and a root/grater visual.

Below are the top-performing teasers. Images used included leaves, nuts, horseradish, and test strips. No AI generators were used — these were archive creatives that had already proven effective. Creative performance was tracked exclusively by EPC.

Here you can see the breakdown by impressions and clicks: 

Campaign Setup

The campaign was launched targeting Mexico, exclusively on mobile devices — desktop was cut from the start to avoid spreading the budget thin. Language, naturally, was Spanish. The minimum test budget was $100, with the click bid set from $0.025. Nothing fancy: the goal was simply to understand how the traffic source performs under standard conditions.

The AdsKeeper manager provided a whitelist of placements for the offer, which made the launch faster than expected. That's another plus — no need to build out zones from scratch or run blind. The landing page used a fixed price of 590 Mexican pesos; no adaptation for the traffic source was needed, it worked as-is.

Campaign Launch, Optimization, and Scaling

Optimization was handled at the source level using auto-rules: if the conversion cost exceeded $6 over a 7-day period, the placement was blocked.

Bid rates (CPC bids) across placements were also adjusted manually. Those delivering cheap leads with no approval were switched off. Placements were passed to the affiliate program for analysis — the focus was not only on lead cost, but also on approval rate and EPC, since some placements generated cheap leads that failed moderation.

As for scaling — the entire pour was primarily aimed at testing the source and understanding how everything works here, so the case stayed within the scope of a single "lazy" test.

Over the course of the campaign, $2,661.66 was spent, generating 94,240 clicks and a total of 674 conversions according to AdsKeeper stats at $3.95 each (394 conversions reached the affiliate program, of which 248 were valid, and 128 were approved). EPC — $0.06. Traffic spend: $2,661.66; revenue (per affiliate program stats) — $3,547; net profit came to $885. ROI — 33%. We consider the creative a success.

Takeaways from working with native ads 

Experience has shown that native advertising can work even with a "lazy" launch — without deep configuration. The key is to enter with a solid funnel and not neglect the basics, such as white/blacklists. It's also important to understand that pre-landers can perform very differently depending on the source: what converts well on Facebook may yield zero results in native. So you need to be prepared to test and swap out creatives.

On bids — the optimal approach is to set the bid at 80–90% of the recommended level, then adjust manually. It's important not only to lower bids on expensive placements, but also to raise them on those delivering cheap, high-quality conversions. Native works on auction principles: lower your bid and you get remnant, lower-quality traffic; raise it too high and you risk overpaying and going into the red.

That's why it's critical to track the balance between CPC and CTR: a teaser with a low CTR may deliver a cheap conversion, but requires a higher bid to break into impressions. The takeaway: bid management across placements and creatives needs to be "live," and manual adjustments often outperform automated rules.