Tony Dinh spent $2,052 on LinkedIn Ads for TypingMind Custom and stopped: the targeting couldn't reach his buyers
- 183,663
- Impressions
- 337
- Clicks (round 1, $300)
- $2,052
- Budget
- Not disclosed
- Duration
Tony Dinh tested LinkedIn Ads for TypingMind Custom in three rounds: broad targeting, Lead Gen forms, then a Lead Gen form turned into an AI quiz. $2,052.03 bought 183,663 impressions and a handful of leads, none of whom had come back when he posted. He stopped the channel.
01The problem
He wanted more sign-ups for TypingMind Custom, a tool for pro users who already know AI models, prompts and integrations.
02The hypothesis
If LinkedIn Ads can reach tech business owners, they can buy sign-ups for a pro AI tool.
03The experiment
Round 1 (his first $300): targeted general business owners in tech. 337 clicks, but 95% came from Android phones and most left within 2 seconds. His usual organic traffic is iOS/macOS.
Round 2 (the next $1,000): LinkedIn Ads can't target macOS or desktop, so he switched to Lead Gen forms, where people submit their details inside LinkedIn. 4 leads at about $150 each. He created an account for each lead by hand and emailed them; nobody had replied. The prospects expected a simple AI bot builder, not a tool for people who already know AI well.
Round 3 (the last $600): he turned the Lead Gen form into an AI-knowledge quiz to filter for people who understand AI. $286.17 brought 4 quiz leads ($71.54 each), and a further ~$250 campaign brought 1 more. People took the quiz for fun but weren't interested in the product.
04Results
| Metric | Result |
|---|---|
| ImpressionsLinkedIn Campaign Manager screenshot, all 3 campaign groups | 183,663 |
| Clicks (round 1, $300)95% from Android phones | 337 |
| Lead Gen leads (round 2, $1,000)stated as “~$150 per lead” | 4 |
| Quiz leads ($286.17)$71.54 per lead (screenshot); a further ~$250 campaign got 1 more | 4 |
| Customers“So far, no one has gotten back” (Lead Gen leads) | Not disclosed |
| Spend | $2,052 |
“Not disclosed” means the source doesn't say. We never estimate or fill in missing numbers.
05What worked — and what didn't
What worked
- The quiz version of the Lead Gen form got leads cheaper ($71.54 each vs about $150).
What didn't work
- 95% of clicks from broad targeting came from Android phones and most left within 2 seconds.
- LinkedIn Ads can't target macOS or desktop users.
- Lead Gen leads expected an easy AI bot builder; none replied after he set up their accounts.
- Quiz takers liked the quiz but weren't interested in the product.
06Lessons
Targeting by job title and industry can't find “pro users who already use GPT-4/Claude”. He guessed a general consumer product that anyone can use would do better with ads.
“I think the targeting capability of LinkedIn Ads is not enough for me to reach my target audience.”
07Source & evidence
- View original
Originally shared on X · Apr 29, 2024
LinkedIn Ads experiment
Public source. Extracted from publicly available content by the founder.
Want to try this?
Copy this experiment
A checklist built from what Tony Dinh actually did. Their numbers are a reference, not a promise.
- 1
Pick the channel and tactic
LinkedIn Ads — Broad targeting → Lead Gen forms → quiz Lead Gen form.
- 2
Set a budget cap
Tony Dinh spent $2,052. Decide your own cap before you start.
- 3
Set an end date
Tony Dinh didn't say how long it ran. Pick a review date before you start.
- 4
Copy what worked
From Tony Dinh's experience:
- ☐The quiz version of the Lead Gen form got leads cheaper ($71.54 each vs about $150).
- 5
Avoid what didn't
Where it went wrong:
- ☐95% of clicks from broad targeting came from Android phones and most left within 2 seconds.
- ☐LinkedIn Ads can't target macOS or desktop users.
- ☐Lead Gen leads expected an easy AI bot builder; none replied after he set up their accounts.
- ☐Quiz takers liked the quiz but weren't interested in the product.
- 6
Track the same numbers
So you can compare like for like:
- ☐Impressions
- ☐Clicks (round 1, $300)
- ☐Lead Gen leads (round 2, $1,000)
- ☐Quiz leads ($286.17)
- ☐Customers
- 7
Compare, then share what happened
Tony Dinh's headline result: 183,663 impressions. Yours will differ — that's the point. Win or flop, the result is worth sharing.
Was this experiment useful?
Want the full playbook behind this?
Tony Dinh hasn't published a step-by-step playbook. If enough readers ask, we'll ask Tony Dinh for one.
Related experiments
- 164
- sales
- $4.2K
- revenue
- 6.2K
- website visits
Xnapper's Product Hunt launch: #1 of the day, 164 sales, $4,212 — driven by a beta email list
Tony Dinh · Xnapper
Tony Dinh launched Xnapper, a screenshot tool, on Product Hunt and hit #1 Product of the Day: 1,428 votes, 6.2K website visits, 164 sales and $4,212. He credits the launch-day email to his beta list most.
- 23K
- google impressions per day, after
- 52.7K
- clicks, whole chart
- 1.8%
- average click rate, whole chart
Tibo ran his SEO tool Outrank on SuperX: Google impressions went from about 2K a day to 23K
Tibo · SuperX
Tibo turned on Outrank — his own AI SEO tool — for SuperX in November 2025. It did the keyword research, wrote the articles, got contextual backlinks and built some free tools. By September 2026 daily Google impressions had gone from about 2K to 23K. Signups and revenue weren't shared.
- 3
- paying customers
- $201
- first MRR
- 80
- DMs sent
Pratham Naik sent 80 LinkedIn DMs in a week: 12 replies, 8 users, 3 paying ($201 MRR)
Pratham Naik · BearConnect
Instead of building an X audience or launching on Product Hunt, Pratham Naik DMed 80 agency owners from LinkedIn groups in one week, offering free access for honest feedback. 12 replied, 8 used BearConnect, and 3 paid within two weeks: his first $201 MRR.