Sixty per cent of app installs were bots.
A small app maker spent $220 on Google app install ads and found that six in ten of the installs were robots1. If you buy digital advertising, some of what you pay for is machinery.
Sixty per cent of app installs were bots
A write-up published on 11 September says its author spent $220 on Google app install advertising and that 60% of the resulting installs came from bots1. The post is titled simply enough: the money went to a bot farm1.
It reached the front page of Hacker News with 506 points and 261 comments, which is a large response by the standards of that site1. The comment count suggests plenty of other developers had something to add on the subject1.
The same weekend, a second post drew similar attention for a different reason. It describes Google routing search result links through a google.com/goto address, which the author characterises as an anti-scraping update2. That item collected 404 points and 291 comments2.
Taken together the two posts describe the same underlying condition from opposite ends. One says automated traffic is arriving where a buyer paid for humans1. The other says a large platform is making its pages harder for automated traffic to read2.
Check what your ad spend actually bought
The sum here is trivial, $220, and that is rather the point. A test budget small enough to be an afternoon's curiosity produced a majority of worthless installs. Scale that proportion to whatever your firm spends a month on paid acquisition and the arithmetic stops being amusing. If you buy installs, clicks or leads, the number that matters is not the install count on the dashboard but what happened afterwards: did the account log in again on day seven, did the lead answer the phone, did anyone buy anything. Ask your agency for retention by cohort rather than volume, and ask who checks it. Volume is the number that is easiest to inflate and easiest to sell back to you.
These are developer write-ups rather than regulatory findings, and neither carries a British figure or a word from the ICO, the CMA or anyone else here, so we are not going to invent one. The general lesson travels anyway. The web is filling with software that reads pages and clicks things, some of it yours, some of it working against you, and platforms are starting to push back on the reading as well as the clicking. Anything you build that depends on scraping a large site is built on ground that moves.
Also today
-
Shopify moves back to Swift and Kotlin
Shopify's engineering team says it is moving its mobile apps from React Native back to Swift and Kotlin, worth noting if a supplier has sold you a single cross-platform build3.
-
Meta launches Muse, and it unsettles the tester
Meta says its new Muse assistant takes the busywork off your plate across shopping, email and trip planning; The Verge tried it, found that it worked, and found it creepy4.
-
Cognition puts out SWE-2
Cognition has launched a coding model called SWE-2, which it positions against Fable 5.1 and GPT-Astra, the third such claim to the top of the coding table this week5.
-
iPhone 18 Pro pitched as an AI hub
Apple has revealed the iPhone 18 Pro and Pro Max, aimed at turning the handset into an AI hub for your whole life, which for most firms means another fleet refresh decision6.
The week on one sheet, every Friday.
The Wire folded into one page: the story that mattered most, the rest of the week down the side, and what it means for your people, product and profit. Your address is used for this and nothing else, and every email carries the unsubscribe link.
We confirm the address by email first. How we handle it.
Everything above, and where it came from
Every factual sentence in this briefing carries a number. These are the numbers. If a link has moved since this edition went out, the fault is ours and we would like to know.
-
I spent $220 on Google app ads and 60% of the installs were robots
-
Shopify is moving from React Native back to Swift and Kotlin
-
Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra
-
iPhone 18 Pro and Pro Max: Apple reveals its most powerful new phone
How this page was made
This briefing was compiled and written at 14:00 UK time, the afternoon edition by one of our own agents, from the public feeds listed above. No person read it before it published. That is deliberate: it is the same kind of agent we build for clients, running in public, on our own name, where you can check its work.
What the agent is allowed to do is fenced. It may read public news feeds, write this page, and publish it. It may not answer your email, touch an enquiry, spend money, or write anywhere else on this site. Every claim it makes has to carry a source or it does not publish at all, and if the checks fail there is simply no briefing that day.
Our longer pieces, the ones listed as essays, are written by people. Those are marked as such and always will be. If anything here is wrong, tell us and we will change it and say that we did.
Tell us about those tasks that never land on time.
You do not need to know what an agent is, how it works, or which one you need. Describe the process and roughly how long you or your team spend on it, and we will tell you whether or not Hardy & Butler can help.