Are you cutting traffic too fast and losing profitable sources before they have time to convert? This Tango Live case study shows how a broader testing approach helped scale 1.3M+ Mobile Popunder impressions – without over-optimizing the campaign.
Performance marketers are trained to optimize fast. A source spends without converting? Cut it. One OS version looks weaker? Remove it. A GEO starts behind the others? Move the budget.
Sometimes that is exactly the right call. And sometimes the smarter move is to give the traffic enough time to show what it can actually do.
That was the case with Tango Live. The advertiser had a strong mobile product, a funnel already adapted for the target markets, and a clear understanding of what mattered after the first conversion. Instead of narrowing the campaign immediately, the team started broad, let the full HilltopAds feed collect data, and only made a few targeted changes once there was enough signal to work with.
Have a mobile-first offer to test?
Sign up with HilltopAds and put Mobile Popunder traffic to work.
Over the selected 6-day period, the Mobile Popunder campaign delivered more than 1.3 million impressions across Saudi Arabia, the UAE, and Turkey at a $0.85 CPM.
The interesting part, though, was not the amount of traffic. It was how little needed to be changed once the right product met the right traffic source.
Key Points
Offer: Tango Live
GEO: Saudi Arabia (SA), United Arab Emirates (AE), Turkey (TR)
Traffic: Non-mainstream High and Medium Activity
Selected campaign period: July 14-19, 2026
Ad format: Popunder Mobile
Pricing model: CPM
CPM: $0.85
Impressions: 1,300,630
Spend: $1,105.54
Why Tango Is More Than a Simple Lead-Gen Offer
Before getting into the campaign, it helps to understand what kind of product Tango actually is. Calling it simply a dating offer misses most of the mechanics that matter for traffic acquisition.
Tango is built around live video and creator interaction. Users can watch broadcasts, chat with creators, send virtual gifts, buy coins, access premium interactions, and subscribe to individual broadcasters. Tango introduced three subscription levels for creator content, including access to exclusive posts and premium broadcasts.
Read our article on how to create a High-Converting Landing Page:
Its monetization does not stop at one payment either. Tango’s current loyalty system is built around continued coin purchases and gifting. Users progress through multiple loyalty levels by buying coins and supporting streamers, while its separate Crowns system explicitly tracks lifetime spending on an account.
- For a simple lead form, the first conversion may tell you most of what you need to know.
- For a product like Tango, the first lead is much closer to the beginning of the commercial relationship.
A user can enter the platform, discover creators, interact, make a first payment, and then keep spending. That is why the advertiser was not looking only at leads in this campaign. Lead cost mattered, but so did the cost of the first deposit – and, on the advertiser’s side, repeat deposits.
Roman
Business Development Manager at HilltopAds
We agreed on target costs for both the lead and the deposit from the start. The advertiser also tracked repeat deposits on their side and shared that data with us.
Those deeper events later became important when the team started deciding which traffic was actually worth keeping.
Why the Mobile Part Matters
One product detail made this especially relevant for the campaign. Tango currently supports full browser access on mobile. For iPhone users outside the EU, Tango says the platform can be accessed through a mobile browser, where users can watch live content, chat with creators, interact in real time, and send gifts.
All three GEOs in this campaign – Saudi Arabia, the UAE, and Turkey – are outside the EU. That does not automatically make every mobile campaign successful, but it removes an obvious point of friction: Tango is not a product that only makes sense after forcing an iPhone user through a completely different desktop experience.
The product itself is already designed to be consumed from the phone. And that is where Mobile Popunder became a particularly natural test.
Register with the best ad network for advertisers HilltopAds
- 273B+ Impressions Worldwide
- Full manager support
- Traffic from 250+ сountries
- Premium traffic sources only
- Minimum deposit $100
Why Tango and Mobile Popunder Made Sense Together
Some offers need several rounds of targeting changes before the right traffic combination appears. This one showed promise almost immediately.
With Mobile Popunder, the user is already browsing on a phone when the advertising flow opens. Tango then continues that experience inside a product centered on video, real-time interaction, and mobile consumption.
There was no need to invent a complicated traffic hypothesis around it. The format and the product simply made sense together.
Roman
Business Development Manager at HilltopAds
Mobile Popunder showed strong results almost immediately. It was one of those cases where the offer and the traffic source were a very good match. The product itself is strong, and the funnel is well adapted for mobile devices.
Pop traffic can generate a lot of volume, but that volume only helps if the experience still makes sense once the user arrives. If the landing experience, payment flow, localization, or product itself creates friction, buying more impressions usually just exposes that problem faster.
That was not the issue here. Tango already had the mobile experience in place, and HilltopAds traffic gave the advertiser enough scale to see how that funnel behaved.
Campaign Setup in HilltopAds
Before you can start an advertising campaign on HilltopAds, you’ll need to sign up as an advertiser. Once your account is ready, here’s how to create a similar Mobile Popunder campaign.
First, go to Manage Campaigns and click Add Campaign.
Start with the basic Campaign settings:
- Ad format – Popunder Mobile
- Pricing model – CPM
- Traffic channels – Non-mainstream High + Medium
The campaign started with both Non-mainstream High and Medium Activity traffic available, giving the advertiser a broader pool of sources to test before making source-level decisions.
Next, configure the Targeting settings.
For this campaign:
- GEO – Saudi Arabia (SA), United Arab Emirates (AE), Turkey (TR)
- Device type – Mobile
- Operating system – iOS
The same core approach was used across all three GEOs. According to Roman, there was no need to build a separate setup for each market because the Tango funnel was already well adapted to the target audience.
At this stage, the important thing was to collect enough traffic before making narrower decisions inside iOS.
The final selection of iOS 18 and iOS 26 came later, during optimization, rather than being treated as a fixed assumption from the start.
Then configure the Traffic filters:
- Proxy – Disallow
- WebView – Disallow
For this campaign, both Proxy and WebView traffic were disabled.
This also kept the setup aligned with Tango’s mobile browser experience.
Next comes the CPM bid and budget.
For the campaign analyzed in this case:
- CPM – $0.85
- Daily budget limit – $200
During the selected 6-day period, the campaign generated:
- 1,300,630 impressions
- $1,105.54 in spend
The numbers line up directly with the configured CPM:
1,300,630 × $0.85 / 1,000 = $1,105.54
That works out to roughly 216,772 impressions per day during the analyzed period. The setup itself was not especially complicated, but what happened after launch was more interesting.
Register with HilltopAds and get:
- Advanced targeting options
- Direct traffic sources
- Self-serve platform
- Fully-managed service
- Postback tracking
Then We Did Something Unusual: We Left the Feed Alone
The advertiser gave the campaign something performance teams do not always give new traffic: time. Up to two weeks had been allowed for testing the full feed, which meant HilltopAds did not have to start blacklisting zones after the first handful of conversions.
Roman
Business Development Manager at HilltopAds
The advertiser gave us two weeks to test the full feed. At the beginning, no sources were excluded at all. Once we had data, source optimization was handled with a standard manual blacklist.
The results covered in this case are a six-day snapshot from July 14 through July 19, so this should not be read as a two-week performance report. The important point is that the testing framework itself allowed enough time for sources to collect meaningful data before they were judged.
That patience mattered because the advertiser was not evaluating traffic on one event. A lead could look cheap and never turn into a paying user. Another source could produce fewer initial actions but send users who deposited – and potentially deposited again.
Cutting sources too quickly would have meant judging a deeper funnel using only its earliest signal – a risky shortcut when the value of a user appears later.
A campaign starts spending, the first reports come in, and every source with an ugly-looking first day suddenly feels like waste. But small samples can be deceptive. Our own 2026 guide to profitable traffic makes the same point: breaking data into groups that are too small can give advertisers the wrong impression, and sources should be evaluated after they have spent enough to make the comparison useful.
In this case, the advertiser waited long enough for that comparison to become useful before the team started cutting traffic.
Read our article about alternatives to RichAds:
Only Two Things Earned a Cut
Once the campaign had enough data, the team did optimize it. The difference is that the optimization did not become a rebuild.
Two changes mattered.
Narrowing the OS versions
The campaign was ultimately kept on iOS 18 and iOS 26. Roman also pointed to the spending profile of users on these iOS versions as an attractive part of the audience for this offer.
Roman
Business Development Manager at HilltopAds
The product quality and the mobile funnel were already strong. We also liked the purchasing potential of the audience on the selected iOS versions. The main targeting adjustment was keeping iOS 18 and 26.
It is worth clarifying the version naming here. Apple moved from iOS 18 to the new year-based naming system with iOS 26, which was released in September 2025. So the combination of versions is intentional, not a typo.
Manual source blacklisting
The second adjustment was familiar: sources that proved weaker after collecting enough data were removed manually.
There was no reported need for a completely separate bidding structure by country, aggressive dayparting, constant CPM changes, or a narrow whitelist from day one.
The campaign did not need a full rebuild. It only needed the weaker parts removed once the data was strong enough to support that decision.
Optimization did not create the campaign’s product-traffic fit. It cleaned up the parts that did not deserve to stay once that fit was already visible.
- Register
- with hilltopads
- and get
- 273B+
- of ad impressions!
One Funnel, Three GEOs – but Saudi Arabia Brought the Volume
Saudi Arabia, the UAE, and Turkey were not managed as three completely different campaigns. According to Roman, the product and user funnel were already adapted well enough for the target region that the same core approach could be used across the three GEOs.
There was no need to build a separate traffic strategy for every country just for the sake of localization. One market did stand out on volume, however: Saudi Arabia delivered the highest number of impressions and users.
The broader market context helps explain why SA had enough room to scale. DataReportal’s Digital 2026 report puts internet penetration in Saudi Arabia at 99%, with 34.4 million internet users and 48.7 million active cellular mobile connections at the end of 2025. It also reports that 98.5% of mobile connections were capable of using broadband mobile networks.
And during the exact month of this campaign, iOS had a particularly strong position in the Saudi mobile market. Statcounter recorded iOS at 50.12% of mobile OS usage in Saudi Arabia in July 2026.
That does not mean an iOS campaign automatically converts just because it runs in Saudi Arabia. Market-share data tells us there is an audience; it does not tell us that audience will buy.
In this case, though, there was enough mobile iOS scale for the product-traffic match to work with. Saudi Arabia simply supplied more of it.
For this campaign, the way the advertiser measured value mattered more than another top-line traffic statistic.
From the start, there were two acquisition benchmarks: target cost per lead and target cost per deposit. The advertiser then went a level deeper and tracked repeat deposits internally.
Tango gives users multiple reasons to continue spending after their first interaction. Coins can be purchased through Tango and used across the platform, while subscriptions can renew monthly and the current loyalty system is explicitly designed around ongoing purchasing and gifting behavior.
So in a campaign like this, judging quality only by the cheapest first lead would leave out a large part of the commercial picture. A source that wins on CPL is not automatically the source that wins on downstream value.
That is also why the full-feed test was useful: it gave the advertiser more room to compare traffic against the events that actually mattered to the business.
Exact lead, deposit, and repeat-deposit counts were not provided for publication, so we are not going to reverse-engineer a fake CPA or ROI from incomplete data.
The confirmed result from this period is the traffic side of the equation: 1.3M+ impressions delivered at the planned $0.85 CPM, with Saudi Arabia providing the strongest volume and the campaign continuing to be evaluated against deeper advertiser-side events.
The traffic numbers were only the top layer; the advertiser was still judging quality through deposits and repeat deposits.
We recommend that you study the successful cases of our advertisers:
Why Less Optimization Worked Here
The lesson here is not to launch broad and avoid optimization. It is to give the campaign enough data before deciding what actually needs to change.
Tango already had a strong mobile product. Its monetization extended beyond a first lead. The funnel was adapted for the target markets. Mobile Popunder put users into a journey that made sense on the device they were already using. And the advertiser was willing to give the feed enough time to show which sources were actually valuable.
Only then did the team narrow the setup. The campaign kept iOS 18 and 26. Weak sources moved into a manual blacklist. The rest did not need to be constantly rebuilt.
The campaign did not need constant intervention; it needed a few changes backed by enough data.
For advertisers running live-streaming, social video, subscription, or other products with value deeper in the funnel, that may be the more useful lesson from this Tango campaign: let the traffic collect enough history to tell you what it is worth – then optimize around the business event that actually matters.