Hypertune · Google Ads

Hypertune on Google: what we're testing, what it costs, and how we'll judge it

Leadership report · 1 October 2026

From Vinson · 1 October 2026 · for leadership

Hey team,

As of today Hypertune is advertising on Google for the first time. It's a four-week test of $3,000 across Google search, YouTube and retargeting past visitors, and it ends on its own on 1 November. We'll judge it on one number: what each new paying customer costs, counted in Stripe, not in Google's own reports.

It's set up as a pass or fail test. If Google brings in new paying customers at $12 or less, we scale it. If they cost more than $25, it stops. Four weeks and $3,000 is enough to tell those two apart, but not $12 from $15, so the read is "is this a channel for us or not".

Why Google, and why now:

Buyers decide on the first visit. 79% of new paying customers paid the same day they first arrived, so an ad that reaches someone with a problem today can sell today.
The auction is nearly empty. Of the 89 searches we tested in Google's own planner, 61 have too few advertisers for Google to even estimate a price. Of ten competitors checked, only one (ExitLag, which sells ping tools, not optimisation) runs Google ads.
The economics work at $12. Of July's monthly payers, 66% paid a second time, and they have paid $17.24 each so far. A monthly customer bought at $12 has paid it back by their second payment, a yearly one on day one.

What's running

Campaign	What it does	Four weeks	Per day	Status, 1 October
Search, problem searches	Our ads on 73 searches such as "fps drops", "low fps on high end pc" and "valorant fps"	$1,150	$41	Live, all ads approved
Search, competitor searches	Our ads when people search four competitor names. The ads never name a competitor.	$150	$5	Live, all ads approved
YouTube and retargeting	Our 16-second "same PC, Hypertune off, then on" video on YouTube, to new audiences who search for FPS and PC performance, and to the roughly 1,600 people who visited hypertune.gg in the last 30 days	$1,600	$57	Live tonight, ads in Google's review
Held back	Brand protection if anyone bids on our name, otherwise added to the strongest campaign in week three	$100
Total		$3,000	about $103

Where: the US, Canada, Australia, New Zealand, the UK, Ireland, Denmark, Finland and Sweden. Full-price countries, English ads, and the highest conversion rates in our data.
Computers only. A phone visitor converts at a quarter of the rate of a desktop visitor (0.9% against 3.5%), so no campaign shows on phones.
Banners from the design team join the YouTube and retargeting campaign when they're delivered.

The guardrails

Click prices are capped at $0.40 to $0.60, and never above $0.75, which is break-even at $25 per customer.
Anything that spends without selling is cut automatically. A search term at $25 with no paying customer pauses itself. An audience or video at $100 with none is paused.
Each campaign is read at $500 of spend. Under $12 per paying customer, it scales. $12 to $25, it stays and gets fixed. Over $25, it stops and its money moves to the strongest campaign.
The spend can't run over. Every campaign has a fixed total and an end date of 1 November, and an alert fires if a single day passes $150.

How we'll judge it

Stripe is the judge. Every ad link is tagged, so each new subscription in Stripe shows which campaign it came from. Google's own count only steers its bidding: it sees about 60% of purchases, so Google's reports will look worse than the truth.

Kevin is adding a paid ads section to the admin dashboard, next to the rest of the business: spend against plan, paying customers from Google in Stripe, and cost per customer against the $12 target and $25 stop line. The first version is due Monday 5 October.

Timeline

When	What
1 October	Search live. YouTube and retargeting switched on, ads in review.
Week one	Checked daily: spend, click prices, search terms to exclude, paying customers in Stripe
Every Monday	A five-line update: spend, paying customers, cost per customer by campaign, what changed, what's next
Each campaign at $500	Scale, fix or stop, by the rules above
End of week two	Money moves from anything above $25 to the strongest campaign
1 November	Budgets end automatically
9 November	The verdict, a week after the end so late payments land. Decides the month-two budget, and whether Meta ads start, as agreed, after Google's first read.

Decisions needed

Approve the test budget: $3,000 over four weeks. $1,300 on search (of which $150 on competitor searches), $1,600 on YouTube and retargeting, $100 held back. Austin agreed it on 29 September, and the campaigns stop on their own at these totals on 1 November.
Agree how it's judged. Pass or fail on cost per new paying customer in Stripe: $12 or less passes, over $25 fails, verdict on 9 November. Recommended.
US YouTube around the midterms. If US YouTube prices jump more than 30% after 26 October, that week's YouTube money moves to the other eight countries. Recommended, no change up front.
Month two. Confirm that any spend above the $3,000 comes back to you with the 9 November verdict, with the numbers to support it. Nothing increases before then.

I run the account day to day and send the Monday update.