Being right doesn’t mean you win when building a startup

Hello everyone - we’ve been really quiet recently...

...that’s because, after a wild 18 months, we’re completing a private sale of Ting and closing the company.

To wrap it up, I thought what better than a full-colour, HD play-by-play (or blow-by-blow) of those 18 months. How we got so much right and still got it wrong.

To everyone who upvoted and supported us along the way: it meant more than you’ll know. I hope some of our lessons help you get further than we did...

Three lessons I keep coming back to

First, the lessons, in case you churn before reaching the timeline.

1. Simple and obvious is good. Hard to replicate is better.

Ting always felt obvious to me. Why wouldn’t you want to CC an AI into an email thread and have it manage the repetitive work for you?

The same was true of WhatsApp. Wouldn’t it be useful to text your agent and have it pick up some work?

It wasn’t happening. Then it became inevitable.

VC interest went from red hot to chilled. Someone told me that if you’re building a Google feature, it won’t be long before Google builds it.

They did. So did Superhuman, Read AI and plenty of others.

If I did it again, I’d find something harder to replicate by going deeper, perhaps rethinking the calendar altogether rather than building an agent on top of it.

2. Pay yourself properly.

There is already so much going wrong every day as a founder. You’re earning far less than you did before, with none of the stability.

I decided to pay myself the least on the team so we could put as much money as possible into the product and the people building it.

That sounds honourable. In reality, you’re adding more pressure to your life at exactly the moment your stress is reaching new peaks.

Pay yourself a reasonable amount.

3. Focus on truth-seeking.

When you’re inside a large company, it’s easy to abstract the truth.

A campaign report lands and you find the best angle for your boss. Your boss finds the best angle for theirs. By the time it reaches the founder, it’s all noise and no signal.

As a founder, you need the truth because every day counts. You have to be able to say openly how well or badly things are going.

Build a culture where mistakes and losses are discussed and owned quickly. You’ll find a new speed.

You can never go wrong reporting the truth. You can go very wrong reporting an abstraction.

The 18-month timeline

January 2025

I got back from a run and found four emails about rescheduling different meetings that day.

Thirty minutes later, I thought: wouldn’t it be useful if I could CC an AI into these threads and let it handle everything?

February

I couldn’t shake the idea.

I called it Ting, as a play on “meeting”: Meet-Ting. Once an idea has a name, you know it’s got deep into your subconscious.

I messaged Ol, who I knew from TikTok, for a vibe check. I didn't know he was launching a venture studio. After hearing the idea, he said: “If you build it, I’ll invest.”

A DM can change everything.

March

I handed in my notice.

We started building during weekends and evenings. We spoke to developers, agencies and consultants before creating a prototype.

It worked. It wasn’t yet delightful, but it worked.

Because we had vibe-coded it, every change broke something else.

We had to rebuild. A lot.

April

I speculatively pitched a16z Speedrun.

We reached the final interview and were told we’d hear back about a $1 million investment.

Imagine how crazy that felt.

May

We planned a three-week fundraising window and filled 75% of the round.

Then a major UK VC offered to take the whole thing. We decided to stand by our angels and finish the round with them instead.

From a relationship standpoint, I think we made the right decision.

For the business, I’m less sure. A VC is incentivised to get you to the next round and bring their VC network. A potentially pivotal decision in hindsight.

June

I went full-time.

a16z decided not to invest. They thought we had underestimated the technical complexity.

I already knew we needed an exceptional technical co-founder. Through the YC ecosystem, I met Mariana: an applied maths genius who had been working in AI before it was cool.

We convinced her to join.

July

We launched here on Product Hunt and finished sixth, despite starting from zero and launching on the same day as OpenAI and several established products.

The launch created real momentum. A massive PC manufacturer's innovation team (under NDA, sigh) approached us about an internal rollout and possible deeper integration.

After the low of a16z, this was a new high.

August

We expanded beyond a Gmail-only beta and removed the waitlist.

I had resisted opening up because Ting still occasionally made mistakes. I worried users would be upset.

I was wrong. Restricting access was restricting our signal.

We quickly doubled from 100 users to 200.

Every user had to connect both their calendar and email, so this was a high-trust product. We were happy with the growth.

September

Said 'largest PC manufacturer' went through a reorganisation.

The opportunity disappeared. That was that.

We kept building and reached 400 users.

October

A large AI community discovered Ting. An Instagram post caused a huge spike in traffic and brought a wave of VCs to us.

We took the calls but told them we weren’t raising.

Another lesson: always be ready to raise, especially as an AI startup. These conversations can take months, and investor interest doesn’t wait for your fundraising window.

November

Our growth convinced us to start talking to VCs properly.

Mariana and I went to San Francisco for a ridiculous week of meetings with tier-one investors and Microsoft.

The conversations were good. Nothing became concrete.

We decided to focus on markets where we had an edge. AI adoption was accelerating in Brazil, and we had good connections to the market: I speak Portuguese, Mariana is Portuguese and my wife is Brazilian.

We ran a test in Portuguese.

We recruited 100s of Brazilian users in three weeks, mainly thanks to Ting on WhatsApp. IYKYK.

December

Based on competitors’ communications, social activity and web traffic, Ting appeared to have become the most-used scheduling agent of its kind.

But the category was filling up. Sequoia had backed a competitor. Another had come through YC. More were appearing every week.

I started contacting their founders.

One of the best decisions I made.

Speak to the founders building what you’re building. You have a lot in common. You love the same problem in a way very few people ever will.

January 2026

I cold-messaged Tope, the founder of Calendly, and asked if he wanted to meet.

I had studied Calendly obsessively and jokingly thought I might know more about it than he did.

Our conversation was one of the best of my life.

Annoying to really like your category leader. You’re supposed to want to destroy them.

February

We started raising with 1,000 active users, 10% of them using Ting frequently, 5% booking at least 20 meetings p/week with Ting, and 5,000 sign-ups generated from £25,000 of marketing.

We thought that was good.

But investors now wanted to see a path from zero to $10 million in ARR almost immediately. Because we had waited to charge, our revenue story was difficult to make convincing.

We relaunched on Product Hunt, finished in the top three and switched on payments.

I scheduled a call with every one of our first customers, and missed the first one because I was pulled into a technical issue.

I’ll never forget the embarrassment of missing a call with one of the first people willing to pay for the thing we had built, then having to ask them to reschedule. Perhaps that's poetic for a scheduling AI agent?

March

We kept reaching investment committees. Then someone would lose confidence or find a reason to say no.

Several VCs told us that enterprise adoption would help de-risk the investment, so we found two enormous companies willing to become design partners.

We went back to the VCs.

Still not enough.

April

I realised we were building for investors and losing what had made Ting special.

We began exploring a deeper idea: the calendar itself was the root of the problem.

Calendars treat time as binary: free or busy. Real life doesn’t work like that. Time is layered.

You’ll move a meeting if your boss asks. You’ll make room for a deadline. You’ll be more flexible for someone busier than you. A blank hour doesn’t mean every possible use of it has equal value.

Mariana had an idea for an API that could understand and rank those layers.

We started building it.

May

VCs liked the new idea but wanted to see it working.

I struggled to find the same blind faith I’d had in the original Ting. After an honest team conversation, we decided to call time on the company.

I started talking to founders, companies and potential acquirers.

We entered serious discussions with one of the fastest-growing ATS companies in the world. They liked us, but couldn’t justify buying the code because so much of it had been AI-generated.

Our domain expertise was worth more than the company’s assets.

When code is commoditised, is that now the norm?

Eventually, I found a buyer.

June

I began preparing the business to wind down.

I had been self-funding it for several months, so we radically reduced costs. I realised just how much money disappears into all the SaaS tools required to build a product.

The private sale continued working its way through.

Now

I’m finally at peace with the decision and the last 18 months.

You can be so right and still not win. Building a company is random. It’s luck, timing, network and determination.

We predicted where the market was going, grew, found paying customers and eventually sold the company.

But we couldn’t make it through to the next gear.

The experience has prepared me for whatever comes next. It also helped me build a network that was only possible because I threw myself into it.

I can solve difficult problems. I can build and run a team. I can run a business.

I finally understand tax, too.

Should you build? Absolutely.

Just learn from my mistakes first.

What's next? Touching some grass and finding another hard problem with more smart people.

// Dan, Ting + Team