What Makes a Media Buyer Launch a Test in Your Self-Service Ad Platform

When an ad network thinks about media buyers, it usually thinks about the funnel: how to attract them, how to onboard them, how to get the first deposit. Much less attention goes to what the buyer does on the platform the day after launch, and that is exactly where the decision to stay or leave is made.

As part of market research for Rinku, a tracker I am building for media buyers, I talked to people who buy traffic every day: push, pop, banner and native, in ad networks and through their own trackers. I asked them to walk me through a normal working day, and then I watched them work in demos.

What surprised me was how similar the days were. Different people, different verticals, different trackers, and almost the same routine. That routine shows very precisely which platform features a buyer values.

The Loop a Buyer Runs Every Day

A buyer who launches campaigns in networks on their own, rather than through a manager, repeats the same cycle every day.

It starts with costs. Prices in the ad network and in the buyer's tracker can differ, so the first job is to bring spend from the network cabinet into the tracker. Some networks pass costs automatically. Where they don't, buyers move them by hand: export a CSV from the network and import it into their tracker. Without this step ROI in the tracker is fiction, and any optimization is done blind.

Then comes the big picture: yesterday against the day before, against three days ago, against the week. The goal is simple, to understand what changed and where something dropped before opening any specific campaign.

Then priorities. Buyers sort campaigns by spend. A campaign with great profit and small spend can wait, while a campaign burning money at a loss can't. When there are hundreds of active campaigns, colors and quick filters are the only way not to get lost.

After that, the drill-down: network, campaign, zone, slices. And the zone, the source of traffic inside the network, the place where the ad is actually shown, sits at the center of the whole routine. First the buyer looks for zones that burned budget and brought no conversions, and those go to the blacklist. Then the buyer looks for zones that did bring conversions, and raises the bid on them to buy more.

The last step is the one buyers describe with the most irritation: applying the decision. A blacklist in the tracker stops nothing. The zone keeps spending money until someone opens the network cabinet and adds it to the list. One buyer called it monotonous work: you walk from network to network adding black and white lists.

There is also work with creatives, tests, and notes on what was changed. But the core, in the buyers' own words, is analytics of sources inside a network: what works for this campaign and what doesn't.

One important detail is that the loop repeats every day and nobody can run it for the buyer. The analytics belong to them, because only they know what a conversion is worth. According to one buyer, the first pass through a single network takes 20 to 30 minutes. Multiply that by the number of networks and the number of days.

Five Things a Buyer Looks For

1. Transparency of volume and prices before launch

Before configuring anything, a buyer wants to know whether the game is worth the candle. How much traffic is there for this geo, in this format, with this targeting? How much does it cost? Setup takes time: a source in the tracker, an offer, creatives, a campaign. If the volume isn't there, all that time and effort goes to waste.

In my experience, networks that show volume and a price range before the campaign is created save the buyer from a test that was doomed from the start, and save themselves from a campaign that will never spend anything.

2. Auto-rules and goal-price models

If the daily loop has one most repetitive step, it is finding zones that burn budget and sending them to the blacklist. Auto-rules automate exactly that: you set a group of conditions, and when a zone meets them, it is stopped. Buyers gave me a simple example of a rule used in a network: a zone has spent an amount equal to two conversions and brought none, so it goes to the blacklist.

Two things from the research are worth remembering.

First, rules are useful already at the stage of initial optimization. A simple rule protects the test budget, so you don't accidentally spend all of it on one zone. It is also a tool for spreading the budget across as many zones as possible and getting more information about each of them. In my view, rules should be treated as a way to explore the network's inventory, not as blind blacklisting. It is like a pie: you try small slices of many flavors, and the one you like you take whole.

Second, rules need to be re-checked. A zone that was removed long ago may still be sending conversions, and by the end of the month it can turn out to be one of the best. Zone quality floats. A good tool lets the buyer not only cut, but also review what was cut.

In the same group I would put pricing models aimed at the price of the target action, such as CPA Goal, Smart CPM and similar. They move part of the work "by zones" into the bidding logic of the network itself: the platform tries to buy traffic at the conversion price the buyer needs.

3. Dedicated bid per zone

Rules find good and bad zones. But what do you do with a good one? You scale it, which means raising the bid and buying more.

If a bid can only be set at the campaign level, you get unpleasant arithmetic. You raise the bid for one zone and pay more for all the traffic in the campaign, including the zones that don't convert. The way out is to create a separate campaign for that zone, then another one for the next price, and then all of it has to be moderated, updated and not forgotten.

I saw this in practice at AFFMY: buyers created several whitelist campaigns distributed by price, just to move a bid by a couple of cents on a specific zone. When we let buyers set a separate bid per zone inside the same campaign, about 2% of campaigns used it. They generated 3 to 5% of platform revenue, EPC on them grew by 23%, and the total revenue growth was around 4.7%. Part of the effect came from the feature being used in both directions: lowering the bid to save a borderline zone and raising it to buy more on a strong one. I broke this case down in detail in a separate article.

This is functionality that experienced buyers use actively so as not to multiply campaigns. It doesn't slow the work down, and if you use the API, the whole process goes even faster.

4. API access

Experienced buyers rarely work in a single network. They buy traffic in several and use their own tracker as the one place where everything is compared. For them the network cabinet is a place to visit as little as possible.

An API helps a lot here. Through it, costs land in the tracker without manual uploads. That alone takes a load off: going through every network, exporting spend and loading it into your own system takes time, and it is one of the most important actions of the loop described above, where it is better not to make mistakes.

There are other benefits too: blacklists and bid changes can be applied from the tracker, campaigns can be managed (stopped and restarted), and statistics are pulled on request. These are all frequent manual actions, and the ability to reduce them to a single request will definitely interest a buyer in an ad network.

5. Targeting flexibility

If a buyer can't target a narrow segment, you force them to buy traffic on broad targeting. In general that is fine at the start, when the goal is to find a working slice of the network's traffic, but later the targeting has to be narrowed to the one that works. And here the standard nightmare of any network kicks in: targeting cuts the traffic, there is little of it, and spend is low. But in reality you are giving clients a way to buy the most relevant traffic, they are ready to pay a higher bid for it, and what doesn't suit them will be distributed through the rest of the system.

In the interviews this looked very practical: browsers, languages, operating systems, versions, geo, wherever the network allows it. Buyers note that different networks have different targeting settings, and a good buyer builds the test around what each network can give. In a network with limited targeting options, they have to compensate with a long blacklist.

But That Is Not What They Complain About

If you ask a media buyer what is wrong with a network, you will rarely hear these five points. You will hear something simple, like "conversion is too expensive" or "there are no results at all." And partly that is true, because good traffic can't be replaced with anything. So we take it for the truth and decide that the features are secondary.

But look at what an absence of results can be made of. The cost of a conversion depends on how well and how fast the media buyer works on optimization, so it is the result of the daily loop. A zone that eats budget without conversions keeps eating it until the buyer notices, decides and applies the decision in the cabinet. The time between "this zone needs to be stopped" and "this zone is stopped" is paid for with the test budget. The same goes for a zone that converts well but stays on a low bid until someone notices and raises it: the missed volume is paid for with scale.

Every feature on the list helps shorten that reaction time. Auto-rules shorten the time between "bad zone" and "stopped zone." A dedicated bid lets the buyer scale a zone without building a separate campaign around it. The API removes the manual step in the middle. Targeting narrows the delivery down to the relevant audience only. And transparent prices and volumes give a reason to launch new tests when new offers appear.

So when a buyer says "conversion is expensive," behind it there can be a loop that runs slower than it could, on a platform where some of the steps are harder than they need to be. This is my hypothesis based on conversations with buyers: they described the manual pain very clearly, but none of them drew a direct line from "expensive conversion" to "these tools are missing."

What to Do About It If You Run a Network

Think about the day after launch, not only the launch. Onboarding takes one day, while the loop is a repeating process. A platform that makes the loop short gets more of the buyer's budget, because that is where the buyer's time is spent.

Count the manual steps. Take one real buyer and ask them to tell you, or better yet show you, a normal day on your platform. Count how many steps happen outside it: in Excel, in another network's cabinet, in the tracker. Every such step is a reason to choose a competitor who removed it.

Treat the API as a product, not as an integration. For buyers with trackers, and especially with self-written solutions, it decides whether auto-rules, cost sync and blacklists will work at all.

Give them the zone and a price for it. All the decisions in this loop are made at that level. If a zone can't be blacklisted or given a dedicated price, there is a high chance you are losing your most advanced clients.

If you see similar problems in your network, such as buyers leaving quietly or a low conversion from registration to launching active campaigns, let's discuss how to check it and fix it.