5 Painful Lessons For Startups
5 Painful Lessons For Startups
I learned all of them the hard way. Will I do it again? Hell yes.
I haven’t talked much about my life prior to Web3.
But in 2015, I co-founded a healthtech AI startup ($30M raised while I was there), with a vision to improve the health of millions of people using data.
For six years, I spearheaded key business lines focused on mobile consumer healthcare. Building from scratch, our team turned my ideas into actual working products with real users:
A mobile health coaching app where users could converse directly with real coaches
A food scoring algorithm that could tell how healthy your meal was when you snap a photo of it. It used computer vision and machine learning. We licensed the APIs for this tool to other mobile apps
A mobile workplace mental wellness app that tracks and stratifies mental health risk for employees, helping people to stay happy and healthy at work. Coaching & therapy was provided in-app for those who were at high risk of burn out and mental issues.
BUT..
None of these consumer-facing products succeeded despite our best efforts. All of them eventually failed.
But it did provide me with a wealth of insights into tech startups. And first hand experience on the struggles that founders face.
Let me share a few short lessons:
Lesson #1. Distribution Trumps Monetisation
Our initial business model was subscription-based, a necessity given the significant operational costs for our apps, including human health coaches.
We used technology & AI to supercharge our coaches’ efficiency by 20x, but the fixed costs remained. I thought that if we priced our offering at 50% cheaper than than the market rate for coaching consultations, new users would naturally come.
However, we soon discovered that in the crowded health app market, standing out required more than just a great product. It required a clear path to the users.
In other words, distribution takes precedence over monetisation.
We started charging for our product right from the beginning, without sorting out our distribution channels and top-of-funnel approach. Looking back, it was naive of me. Our growth started out at a snail’s pace.
Another reason why having more users is good, even if they’re not paying users: data is often a reasonable moat in tech. The faster you grow and the more data you collect, there is potential to achieve network effects and find the right product feature set quicker.
Lesson #2: The High Cost of B2C Businesses
We targeted the B2C market, but we severely underestimated the resources required to win.
Unless you’re lucky enough to go viral, B2C businesses require large amounts of marketing spend in order to grow and reach escape velocity. This only becomes justifiable if either:
You have lots of funding and prioritising growth at all costs (much harder in today’s funding environment)
The unit economics fundamentally make sense.
Despite having a growth team experimenting with various scrappy tactics, our cost of acquiring customers was high.
Understanding your customer acquisition cost (CAC) and lifetime value (LTV) is crucial to identify opportunities and ensure sustainable growth. CAC payback is one important metric: the number of months required to recover the cost of acquiring a customer.
We once ran a campaign on Facebook that cost us $150+ per new paying user. However, our average user only brought in $60 - $90 over their lifetime. We tried to justify this internally by saying that:
the new users helped us to improve our product faster
there would be economies-of-scale once we reach a certain size
But the negative unit economics was unsustainable and forced us to rethink our strategy.
Lesson #3: B2B2C Businesses Are A Double-Edged Sword
After limited traction in the B2C arena and with no more resources to commit to marketing, we switched track.
We decided to target established businesses who owned the distribution to our target end users: Big Pharma, large insurance companies, hospitals.
The B2B2C concept is enticing: build a consumer app but partner with large enterprises with established channels for distribution — significantly reducing user acquisition costs Clinching one partner could mean millions of users at once.
But in reality, the B2B2C approach is extremely hard to execute well:
You need to build features for both the enterprise customers (backend tracking, customisations) and end users, splitting your product team’s priorities.
You need a high-performance enterprise sales team in addition to an excellent product team. Sales cycles are long and uncertain. Few organisations are built to do both well.
It’s not impossible. Some companies, like Omada Health, have done this very well.
In our case, we partnered with one of the largest insurance companies in Asia-Pacific, integrating our AI tools into their existing member app. This was a win-win, allowing them to offer new features to their users and we were able to monetise our technology without having our own user base.
But soon they wanted more custom features for their backend. This required significant development resources, which took away resources from our consumer-facing product development.
There were consequences: the gap with our competitors widened over time.
Lesson #4: Know The Real Cost of Building an App
Building an app is not a cheap endeavour.
From my experience: a simple yet decent quality native iOS app would cost at least $500K.
This includes not just the development costs, but also the ongoing costs of updates, bug fixes, maintenance, and user support. I’m very skeptical of pitch decks by early stage startups that promise to build out their app with <$300K of capital devoted to development costs.
We had a brilliant engineering and design team (~20 people including PMs, developers, designers) that was capable of building awesome products.
But what’s most important is the product vision: at the end of the day, what should the product do?
We were often torn on which features to prioritise, since there were always too many feature requests, and not enough engineering bandwidth.
Business decisions kept getting in our way: in a bid to grow revenue, I often ended up prioritising features that would help us win a specific partnership deal, rather than features that would have a direct effect on user engagement and retention.
This ultimately led to a decrease in user satisfaction and retention.
Lesson #5: Build Something Cool and the Market Will Come? No.
One of the biggest myths in the startup world is that if you build a really cool product, the users will come.
Our experience showed that this is not the case. Success requires more than just a cool product. It requires a clear understanding of your market, a solid distribution strategy, and a sustainable business model.
It’s Not An Ending
Later, our startup went laser-focused on the B2B enterprise software business for mental health. It has gone on to raise another $45M this year to expand its real-world evidence platform, even as I left to pursue my Web3 interests. I’m still a shareholder, but not involved in its operations anymore.
Will I do this again? Hell yes. The startup journey was a rollercoaster but the lessons learned were worth their weight in Bitcoin.
It has taught me what it takes to succeed and I’ll hold them dearly in my next venture.
If anyone is building a healthcare app that leverages on Web3 tech, let’s talk