Digital advertising is all about staying sharp, making smart moves, and letting data guide the way. Sasha, Head of the CPA department at Adsterra, knows this better than anyone. With over five years of experience under her belt, she’s helped advertisers across all kinds of industries hit their goals with strategies that truly work.
In this interview, Sasha opens up about what makes data-driven media buying so powerful, shares her go-to tips for running successful CPA campaigns, and talks about the trends that are shaping the future of affiliate marketing.
If you’re ready for a fresh perspective from someone who lives and breathes this space, you’re in the right place!
Sasha, could you tell us a little bit more about yourself?
I’ve always been passionate about data, reasoning, and pursuing details. This enthusiasm likely made Adsterra and me a perfect match.
I started my journey here as an Account Manager five years ago and am currently heading the CPA department. As a firm believer in the power of partner care, I’m honing effective user acquisition strategies, putting our partners’ goals first. My team develops tailored media buying approaches for advertisers in various niches.
General Insights and Strategy
How do you define data-driven media buying, and what makes it so powerful?
Generally speaking, when someone adopts a “data-driven marketing approach,” they make strategic decisions based on analyzing and interpreting dynamic data.
When we resort to data-driven media buying, we use data insights and analytics to make informed decisions about purchasing media placements with maximum efficiency. Basically, we’re using real-time, detailed data to target specific audiences, optimize ad placements, and measure effectiveness more precisely.
The data-driven approach allows a media buyer to get insights about precise targeting choices, improves budgeting, and increases audience engagement. I would say that it’s the only way to maximize the output of your ad campaigns.
What are the first steps a media buyer should take when starting a new data-driven campaign?
Any data will depend on how you apply it. So, before diving into figures, set specific objectives for the campaign.
If you’re a product owner, you need to answer the primary question: are you aiming for brand awareness, lead generation, conversions, or customer retention? Clear goals will help you collect the right data and measure success effectively. Choose key performance indicators (KPIs) that align with these goals: cost per acquisition (CPA) or return on ad spend (ROAS).
If you’re an affiliate, a CPA affiliate network or direct advertiser will provide you with all these metrics. You’d just need to adjust your goals to generate profit.
In any case, first, you’ll need to identify and analyze the target audience. Research other media buyers’ experiences promoting the same offers. Invest some time in research—study the forums and case studies. It’s a crucial step to understanding which ad platforms you should allocate your budget to. Case studies usually provide you with some insights into the best traffic sources. Every platform has unique audience targeting capabilities, so aligning platform strengths with audience insights is key.
Then, you use data-driven approach when adding content and creatives. Choose ad research tools, like Meta Ads Library, to prepare the ad creatives tailored to your target audience. A/B testing different creatives can also provide data on which visuals and messages perform best.
And don’t forget to set up real-time analytics–Keitaro is just the right tool for this task.
How do you approach setting KPIs and benchmarks in data-driven media buying campaigns?
As I said, defining the campaign’s primary goal is critical. If we refer to the products advertised on Adsterra, most are promoted to generate revenue rather than to raise awareness or engagement. These are conversion campaigns.
Here at Adsterra, we elevate thousands of businesses, and one of the most popular niches is the App niche: VPNs, antiviruses, and video apps. These offers usually include one specific step in their sales funnel or user flow: an installation, registration, free trial, or subscription.
Can you share 1 example of setting KPIs for app promotion?
Sure!
- Determine the average CPI (cost-per-install) from previous campaigns to set a baseline for what’s acceptable. I strongly recommend asking an account manager of your traffic source or platform for a competitive price for this type of conversion.
- Set a baseline of an acceptable CR (conversion rate) from install to trial and then from trial to paid event. Decide the average price for all of these actions in the funnel.
- If you’re an affiliate getting commission only for paid actions such as paid subscriptions or deposits, but you’re acquiring CPM traffic or paying for less expensive conversions—CPI, CPT—it’s critical to define the acceptable prices for each action in the funnel to ensure you eliminate unprofitable slices in the early stages and avoid overspending.
“A common mistake beginner affiliates would make is increasing the margin too much when defining an acceptable CPA.
It’s natural to think that the greater the margin you take, the more profit you will get. While this seems like a winning strategy, it often leads to traffic drops if you run a CPA campaign. It’s better to have a smaller ROI with larger traffic volumes than a big ROI but no volumes at all.
I enjoy this simple but wise math, so I wrote about boosting traffic for CPA campaigns on Adsterra’s blog. Be sure to check it out!“
Sasha, Head of the CPA department at Adsterra
To sum it up, I advise focusing on data points directly tied to your primary campaign goals and KPIs — such as conversion rate, cost per acquisition, and click-through rate — while filtering out less impactful metrics.
Collecting and Analyzing Data
Which data metrics are most valuable when tracking the success of CPA campaigns?
If we zoom in on Adsterra CPA campaigns, we generally mean campaigns with a fast conversion flow. Installing an app or starting a free trial is considered easy to medium flow, as users perform these actions quite easily. As a result, campaigns generate a stable and consistent eCPM.
We use an eCPM-based auction: the higher and more stable the eCPM, the more traffic you will receive. If the target action is performed too infrequently, there is a high probability that the sources will have a zero eCPM and be removed from the campaign.
Thus, the most valuable metrics for us as your traffic provider are CR, CPA, and, as a result, eCPM—the higher it is, the more successful the campaign is at the auction.
Finally, you need to consider the ratio between installs and your in-app events or other events in the offer’s funnel—prioritize achieving appropriate CPA as your primary goal.
How do you work with segmenting data?
Adsterra aggregates inventory from 35K+ publishers, allowing advertisers to access audiences across a wide variety of websites. We give each ad format on a website an individual number—a Placement ID. There are thousands of those placements in our system, and understandably, it can be overwhelming for a beginner media buyer to handle this data! Honestly, it’s much easier than you might think, and it’s easy to get a hang of it.
Adsterra doesn’t rely on updates of сookies, which is a huge advantage nowadays. Cookies are facing significant challenges due to increasing privacy regulations and shifts in how companies handle user data. Unlike Google and Facebook, ad networks like Adsterra don’t store a user’s personal information, so these big market changes are not affecting us as much as others.
Another aspect of data segmentation embraces creatives: prelanders, visuals, ad copy, and so on. Creatives can highly affect the Impression-to-Install, as well as the Install-to-Purchase conversion rate. I’d suggest having some prelanders ready beforehand. Observing this segment and finding the balance between generating installs and quality leads is crucial.
To sum up, the main segments or clusters media buyers can use in our network are:
- Ad Units – we provide Popunder, Social Bar, Banner, and Native traffic
- Country
- OS and OS versions
- Browsers and their versions
- Placement ID
- Creatives ID / Banner ID
How does a media buyer or affiliate conduct data-driven campaign optimization?
First, make sure you’ve aligned your budget and have room for testing. Always try to save some funds for new tests so they don’t drain your wallet, even in case something goes wrong.
Allow your ad campaigns to be active for some time and collect data about segments we discussed earlier. If you see many red alarming markers in your tracker, don’t let them confuse you. One of the biggest “rookie mistakes” is to stop a campaign that looks unprofitable without analyzing and optimizing.
Then, break down the report from the biggest clusters to the smallest one.
How do you do that?
Break down your offer by:
- Ad unit: is there anything that stands out regarding profitability or losses?
- Country: can you see any clearly unprofitable slices that could be removed straight away from a multi-geo campaign?
- OS: Are there any OS or OS versions showing better results than others? Or are there any versions that are clearly underperforming?
- Prelanders: which one has a bigger CTR, and which one attracts more sales?
The biggest work starts when you get to analyzing the most complex cluster: traffic sources.
While breaking down data by Placement IDs, you can notice some sources that are not bringing many trials while generating a lot of installs. These are the first good candidates for a blacklist.
What can you do with sources with a good install-to-trial ratio but costly sales?
Work on decreasing their rate.
How about sources with a great CPA but smaller traffic volumes?
Increase the rate to get more traffic or shift to the CPM traffic acquisition model. Ask your account manager how to obtain the most traffic from this hot spot.
Depending on the traffic source, you might also notice ad placements that have a suspiciously high install-to-trial ratio. Keep a close eye on them! This behavior is suspicious, and you might need to check with your account manager if they see any weird activity on their side as well.
Overall, my advice for analyzing every cluster and segment is to eliminate the unprofitable ones, work on the almost profitable ones, and scale the profitable ones.
How do you work with creatives effectively and conduct A/B testing?
We at Adsterra highly recommend setting up a rotation of creatives on your side to ensure you get the most transparent results.
For example, you have 4 prelanders promoting your app. Then, your tactic may be setting equal rates and checking, which prelander gives you the highest CR from impression to install. Check if it provides you with good sales, as well!
Keep in mind that some creatives can have really high CTR, although they’re not engaging users to use an app and buy something. Try to find this balance between attracting installs and engaging audiences.
Another vital aspect to consider is behavioral data: look for hours users are most active in the product. As a result of this research, you can choose a time range targeting for your campaigns: this helps you focus ad spending on users most likely to engage or convert.
How often should one check the data?
You need to check the statistics per placement ID at least every other day to achieve the best outcomes. It might sound dreadful, but that’s what most media buyers are doing on a daily basis.
You need to make sure you’re not losing the momentum. Sources come and go; some of them send excellent traffic quality, and some can ruin your statistics.
Advanced Techniques and Emerging Trends
It sounds like a media buyer can’t even go on vacation. Is there a way to take time off in this profession?
Monitoring campaigns every day can be time-consuming if you’re dealing with a lot of data sources, multiple offers, and reports. AI and machine learning transform data-driven media buying, making it more efficient. It looks like you can actually go on vacation and not worry about second-to-second KPI scanning, right?
Whether it relates to bid adjustment or targeting tweaks, AI tools can handle campaigns constantly using dynamic data. And the most apparent automation strategy is creating a script and optimizing your campaigns via API.
What are the API benefits of working with an ad network?
The benefits of using the API of an ad network like Adsterra include a number of advantages in managing, optimizing, and scaling campaigns. Key benefits include the following:
- APIs automate tasks like budget adjustments and bid updates, reducing manual effort and saving time.
- Advertisers can set rules to automatically pause underperforming campaigns or scale up high-performing ones based on real-time reports (blacklisting, whitelisting, and custom bid features).
- APIs provide instant access to performance data, allowing for faster decision-making and campaign optimization.
- Key metrics like impressions, CTR, and CPA can be monitored programmatically, ensuring campaigns remain on track.
- APIs can be integrated into custom dashboards or third-party tools, consolidating multiple ad network metrics into one interface for seamless management.
- APIs ramp up tests of creatives, bids, and targeting settings, which prompts advertisers to optimize campaigns quickly.
APIs of ad networks serve as a powerful competitive advantage to media buyers. They simplify the campaign management process and streamline decision-making. Such a level of control and automation is especially beneficial for businesses that operate broad-reach and complex campaigns.
Combining APIs with AI and machine learning may be a true game changer here. However, this approach requires resorting to third-party developers, as it’s not a “no-code” solution.
How do you become successful using with data-driven media buying?
Affiliate marketing and media buying demand precision, adaptability, and well-balanced decisions. By adopting a data-driven approach, media buyers can optimize costs and rethink their traffic acquisition strategy—literally getting more from the very same targeting.
Adsterra’s ecosystem amplifies your efforts. With them, you’ve got 20+ targeting settings backed up with optimization tools and real-time analytics. Use all of these, applying them to clear KPIs. Only this way will your media buying be data-enriched, not only data-driven.
Most of the data-driven activities stem from strong strategic planning: that’s what I call the future of media buying. You can’t solely rely on data without human expertise. Math and ingenuity come hand in hand!
If you haven’t tried Adsterra yet, you can use promo code F5HWGCD9. Get an extra $50 on a $200+ deposit. The promo code is valid until 08.31.2025.
