Last updated: August 5, 2026
Smith Jones, Performance Media Buyer and Traffic Analyst
I buy paid traffic for a living and I write this site because the public numbers in this category are unreliable in a specific, fixable way: they are copied between directories until nobody remembers which operator page they came from. My working focus is narrow on purpose. Source-level economics: what a thousand impressions costs by country, format and device, what the publisher on the other side of that same impression is paid, and what the gap between those two numbers says about whether a seller is moving real inventory. Everything else on this site follows from that one question, starting with the Cheap Website Traffic page that most readers arrive on.
What I work on
My campaigns run on interruptive inventory rather than search or social: popunder, push notification, in-page push, native and, less often, video pre-roll. Buying happens on self-serve platforms and on exchange supply reached through oRTB feeds. The same offer can therefore be bought at five different prices, depending on which layer of the supply chain it arrives through, and most of my time goes into deciding which layer is worth paying for on a given campaign. The rest is unglamorous: terms pages, payment-method floors, and reconciling what a platform advertises with what it documents. That second half is where this site came from.
My areas of expertise
- Auction mechanics and pricing. The difference between a published bid floor and a real clearing price is where most beginner budgets die. A platform that accepts $0.001 clicks is not a platform that delivers US traffic at $0.001; it is a platform that will happily spend your money on whatever the auction had left over. I read rate cards with that distinction in front of me, and I check the in-dashboard calculators, which move weekly. A blog figure from last year is a historical note, not a rate.
- Billing models. CPC, CPM and CPA behave differently on the same inventory, and on popunder in particular one impression equals one landing page open, which makes CPM directly comparable to a publisher's per-thousand-visit payout. I use that comparison constantly, because it is the fastest way to tell whether a price could be real.
- Invalid traffic. I work with the industry split between general and sophisticated invalid traffic as the IAB and MRC define it, and I check my own inventory two ways: autonomous system numbers to catch datacentre sessions, then IP frequency distribution to catch residential proxy pools that pass the first check cleanly. Both need server or CDN logs, since analytics platforms do not expose the addresses.
- Tracking and attribution. Postbacks, click IDs and platform macros are what turn a campaign report into a source report. If the source ID survives the round trip from click to conversion, optimization happens at source level; if it does not, every later decision rests on campaign averages that hide the placements actually paying for themselves.
- Traffic that is human but worthless. Incentivized clicks, autosurf panels and reward walls pass every bot check because there is a person on the other end. They are diagnosed on conversion behaviour: the volume looks healthy, and the funnel dies at the first step that requires intent. Confusing this category with fraud sends people hunting for bots that were never in the logs.
How I approach a new traffic source
The sequence is the same every time, because deviating from it is how a first budget gets spent on nothing. Postbacks and macros are wired before the first dollar, so conversions come back attached to a publisher source rather than to a campaign. The opening budget is spread thin across many sources instead of concentrated on the cheapest few. That first purchase buys information and nothing else. Only then does log analysis begin: ASN distribution, IP repetition, session shape and engaged time by cohort, judged against what the platform claims to filter on its side.
Bids move last. A source gets cut when it has spent two to three times the target cost per acquisition without converting. That threshold is a decision about money I can afford to lose, not a claim about the source's true conversion rate. A source gets a bid increase after three or more conversions, which is not statistical proof either, only the point where the increase costs less than the information it buys. The reasoning behind those thresholds is set out in how figures are checked.
What I will not publish
No invented campaign results. A sentence like "I withdrew in 26 hours" or "this source returned 240% ROI" is unverifiable by construction, and inventing one would make every checkable number on this site worthless by association. What I can offer is direct observation of documents anyone can open, arithmetic anyone can repeat, and an explicit label on the gaps. The $0.0003 push floor discussed on the home page is the current example: it appears in third-party listings, it could not be confirmed on the operator's own pages, and it is labelled as unconfirmed rather than quietly repeated or quietly dropped. The same discipline governs credentials. There are no degrees or institutional affiliations listed on this page, because listing ones I could not evidence would be the same failure in a different place.
Contact
Questions about a figure, disputes about a number, or a platform parameter that has changed: [email protected], or through the contact page. Mail sent to me is kept for twelve months and nothing else about the visit is stored, as set out in the privacy policy. Professional profile: LinkedIn. The rules governing what gets published here are in the editorial policy.