TensorTonic · Sep 2026 – Dec 2026
Growing an ML practice platform with 52K+ learners
TensorTonic is a platform where people practise machine-learning coding problems. Almost all of its search traffic came from people who already knew its name, and it had no paid channel. In my first two months I launched Google Ads, rewrote how its pages appear in Google and got it into AI answers: organic clicks rose 44% and search click-through rate doubled.
- Company
- TensorTonicML coding-practice platform, 52K+ learners
- My role
- GTM & Growth Engineer Intern
- When
- Sep 2026 – Dec 2026
- Tools
- Google Ads · Google Search Console · SEO/AEO · Apollo · Instagram
- SourceGoogle Search Console, 15 Sep–12 Nov vs prior 59 days
- SourceGoogle Search Console
- 93
- ~60%
The problem
95% of search clicks were branded. People found TensorTonic only when they already knew its name, so search wasn't bringing in new learners. On top of that, there was no paid acquisition channel at all. Growth depended on word of mouth.
To grow, TensorTonic had to show up for people who had never heard of it, both in Google and in the AI tools people now ask for recommendations.
What I did
1. Built a paid channel from zero
I launched five Google Ads search campaigns across the US and India, aimed at two kinds of intent: people comparing TensorTonic's competitors, and people preparing for ML interviews. The ads ran at a 12.2% click-through rate (roughly 1 in 8 people who saw an ad clicked it) and brought in 93 sign-ups in the first 3 weeks.
2. Fixed pages that were seen but not clicked
Some pages appeared often in Google but few people clicked them. I rewrote their titles and meta descriptions, the text people see in search results, so they matched what searchers were looking for.
3. Showed up where people search without the brand name
I published comparison guides and got TensorTonic listed on external platforms. This helps it appear for non-branded searches, and in the answers ChatGPT and Google AI give when someone asks for a tool like it.
4. Closed a partnership
I closed a developer contest with Gradium AI, with a 2M-credit prize pool, to drive engagement. It runs from 20 November to 20 December 2026.
5. Looked after existing customers
I ran a win-back program for about 400 paying subscribers.
6. Opened new channels
I started B2B outbound with Apollo, and Instagram reels.
What I'd do next
Measure what the Gradium contest does for sign-ups and retention once it ends on 20 December, then move ad budget to the campaigns that bring in the most active learners.
Key terms (6): new to the jargon?
- Branded search
- Searches that include the company's name ("tensortonic"). They only come from people who already know you, so they can't grow a business on their own.
- Organic clicks
- Visits from Google's normal (unpaid) results.
- CTR (click-through rate)
- The share of people who saw a result or ad and clicked it. 3.6% means about 36 clicks for every 1,000 times the page appeared.
- SEO / AEO
- Search engine optimisation (ranking in Google) and answer engine optimisation (being mentioned in AI answers from ChatGPT, Google AI and others).
- Search intent
- What the person searching is trying to do, such as comparing tools or preparing for an interview. Ads aimed at a clear intent convert better.
- Win-back
- Outreach to customers who have lapsed or might leave, to bring them back.
What's next