48 of 51 signups came from ads: 30 days inside Plyto's own accounts

48 of 51 signups came from ads

That is the number I keep staring at. We got 51 signups in the last 30 days. 48 of them clicked an ad first. 3 came in without an ad click.

For a company that sells marketing software, that ratio is loud. Our organic, our word of mouth, our LinkedIn posts, our founder DMs, all of it, produced 3 signups. Paid produced 48.

I am not going to pretend that is healthy forever. But it is what happened, and the rest of this post is the honest breakdown of how 48 signups came out of $434.

The spend

Here is the full 30 day picture across our Google Search accounts.

• Oct 30 test, Already running ads: $2 spent, 3,062 impressions, 53 clicks, $0.04 CPC, 1.73% CTR, 0 conversions
• Oct 30 test, Agencies: $0 spent, 59 impressions, 1 click, 1.69% CTR, 0 conversions
• Oct 30 test, In-house marketers: $1 spent, 139 impressions, 3 clicks, $0.39 CPC, 2.16% CTR, 0 conversions
• Oct 30 test, Local service owners: $0 spent, 727 impressions, 4 clicks, 0.55% CTR, 0 conversions
• Plyto campaign, Conversions: $430 spent, 18,763 impressions, 338 clicks, $1.27 CPC, 1.80% CTR, 9 conversions

Totals: $434 spent, 22,750 impressions, 399 clicks, 9 conversions, $1.09 average CPC.

One campaign did almost all the work. The Plyto campaign ate $430 of the $434. The four audience tests together spent $3 and produced 61 clicks and 0 conversions.

The gap between 9 and 48

Google logged 9 conversions. Our own signup count from ad clicks was 48. Same window, same accounts.

That is the thing almost nobody writes about honestly. The conversion count inside Google Ads is not your signup count. Some of those 48 came back on another device, clicked a different ad, did not fire the conversion tag, bounced and returned a week later from a bookmark. Google sees 9. The CRM sees 48. Both are right in their own way, and if you only look at the ads dashboard you will underprice your campaigns by roughly 5x.

Action for tomorrow: match your CRM signup list against your ad click data by email or timestamp before you decide a campaign is broken. If I had killed the Plyto Conversions campaign on 9 conversions and a $48 cost per conversion, I would have killed the thing that brought in 48 signups at about $9 each.

Where the clicks actually landed

Six landing pages took traffic this month. Visits, read past 15 seconds, started typing in a form.

• /: 434 visits, 237 read, 32 started
• /go/chatgpt-ads: 194 visits, 114 read, 30 started
• /go/marketers: 16 visits, 12 read, 1 started
• /go/ai-ads: 15 visits, 8 read, 1 started
• /go/free-crm: 11 visits, 9 read, 1 started
• /go/local: 10 visits, 6 read, 1 started

Two things jump out.

First, /go/chatgpt-ads is the best page we have. 194 visits, 114 read past 15 seconds, 30 started typing. That is a 15% start rate against visits and a 26% start rate against readers. The homepage, by contrast, got 434 visits and only 32 form starts, a 7% start rate against visits.

The chatgpt-ads page is narrower. It talks to one person: someone who just searched something about running ads with ChatGPT. The copy on the page echoes the words in the ad and the words in the search. That page converts at more than double the home page rate.

Second, four of the six landing pages are basically dead. /go/marketers, /go/ai-ads, /go/free-crm, and /go/local combined for 52 visits and 4 form starts. They are not getting enough traffic to judge. We either feed them or kill them. Running six landing pages with this little budget spreads the data too thin to learn anything.

What I am changing this week

Three things, concrete.

1. Push more budget at the shape of traffic hitting /go/chatgpt-ads. That page is already proving the pattern: tight search intent, matching page copy, high read rate, high form start rate. I would rather have two pages at 200 visits than six pages at 30.

2. Pull the four audience test campaigns. They spent $3 and produced no conversions. That is not a failure, it is an inconclusive test. Rolling them over another 30 days at the same spend will produce another inconclusive test. Either fund them to a decision or cut them.

3. Stop reporting on Google's conversion number in isolation. Internally we now track two numbers side by side: conversions Google saw, and signups the CRM attributed to an ad click. This month it was 9 and 48. Next month it will be something else, and the ratio itself is a signal.

The one honest caveat

$434 and 30 days is a small sample. 9 conversions is a small sample. I am writing this because the numbers are specific and the lessons are cheap, not because this is a case study you should copy move for move.

If you take one thing: the page that matches the search wins. /go/chatgpt-ads beat the homepage on every rate that matters, with worse traffic volume. That is not an AI trick. That is just writing the ad's words back on the page people land on.