Smith Jones, Ad Tech Analyst

I have spent the last ten years on both sides of programmatic display: buying pop and push inventory against performance targets, and sitting on the supply side watching the same inventory being sold. That double exposure is the reason I write about the spread between what a buyer pays and what a publisher receives, a subject most coverage of this market avoids because it is uncomfortable for everyone selling into it.

I write the analysis on this site, starting with the working page on Popunder Ads, and my name is on it because anonymous advice about how to spend money is worth what you pay for it.

The Working Day

My working day is mostly source-level analysis. On pop inventory the difference between a campaign that returns and one that quietly drains a budget is rarely the headline bid; it is whether the buyer separated the traffic into individual source identifiers, gave the converting ones a higher price and held everything else at floor. I have spent more hours than I would like to total building and pruning those lists.

The second half of the job is policy. Advertising formats do not die because they stop working, they die because a browser or a platform decides they are abusive. Knowing that Google's publisher rules follow the site rather than the ad window, or that Chrome strips all advertising from a property after thirty days in a failing state, is the difference between a monetisation plan and an expensive surprise. I read those documents when they change, which is more often than the industry notices.

Background

I studied statistics and economics at university, which turned out to be more useful in this field than a marketing degree would have been, mainly because most of the job is deciding whether a number means anything. I moved into digital advertising in 2016 through media buying, spent four years on the demand side across gaming, dating and utility offers, then moved to supply-side analysis where the same inventory looks entirely different.

Along the way I have evaluated somewhere north of forty pop and push platforms in enough depth to have an opinion about them, which is a smaller number than the industry norm of claiming to have tested hundreds. Testing a platform properly means funding it, running traffic long enough for source data to accumulate, and taking a payout. Anything less is reading a features page.

Four Questions Before Any Deposit

Four questions, in order. Can I see real bids before I deposit, or does the platform hide pricing behind a funded account? Can I set a price per traffic source, or is my only lever a single campaign bid? Are there filters that address the specific ways this inventory is faked, meaning proxy exclusion and a considered position on IPv6? Are payout terms written down, or do they turn out to be individual once you ask?

Those four separate platforms more reliably than volume claims do, and none of them appear on a features page. I apply them to every platform I write about, including ones the site has a commercial relationship with, and I publish the failures. On the platform we link to most often, three of five criteria pass and two do not, and the two that do not are named on the page.

What Never Appears Here

I do not invent measurements. If I have not run a campaign myself, I will not write that I withdrew in twenty-six hours or that a source converted at a specific rate. A precise number a reader can check is the worst thing to fabricate, because it is the one thing that destroys everything else on the page when it turns out to be wrong.

I do not write ranked lists of networks with scores. I have never seen one whose ordering was not explicable by commission rates, and I am not going to add another. I also do not write that a platform is the best at anything, because "best" is a claim about a comparison I have not run under controlled conditions.

When I cannot verify something, the page says I could not verify it. There are sentences on this site that read "I found no independent measurement to check this against", and they stay because a gap you are told about is more useful than a gap that has been filled with a plausible guess.

Corrections

I get things wrong. When that happens I want to know quickly, and the fastest route is the contact page with the page name and what specifically is inaccurate. Corrections to facts are made within two working days and dated on the page, and the correction is logged rather than quietly overwritten, so you can see what changed and when. That process is set out in full in the editorial policy.

I am reachable on LinkedIn for professional enquiries, and you can read more about how this site is run and funded on the about page.