Annual Report FY23-24: Nirah Webhooks

Nirah Webhooks

As I mentioned in last year's annual report, I left Setu to work on Nirah webhooks. As a part of Setu, I had the unique experience of building Reverse Penny Drop product whose business metrics depended on webhooks. I knew the complexities and requirements of webhooks to deliver four 9s of reliability, and I could help companies save time on this.

I have stopped working on that. I didn't feel any market pull for this idea. I seem to have been seduced by an "easy idea" and realized the perils of pursuing an easy idea through experience.

Incremental product instead of a game-changing product

The product that I was building was fundamentally incremental. Almost all companies sending webhooks were used to doing it independently and had, over time, realized the importance of observability, reliability, and security. The very best of them had already implemented the things I was offering. There was not much room to make it a game-changing product.

Many of them didn't implement HMAC (though some did), many didn't have observability dashboards for webhook payloads, some of them had bulk flows that required implementing rate limits so as not to overload their receivers servers, manual retries were something which they needed at certain times, but then it wasn't required at all times.

I wouldn't call the features I was offering particularly hard to implement, but I would consider them time-consuming, which was the value proposition. When I was at Setu, I wished someone else had taken care of the webhooks part so that I could focus more on shipping valuable features that could improve the company's competitive advantage.

Perception v/s Reality

People often say, "Perception is a reality in business". This year, I felt it first-hand. I have taken feedback from many of my friends and prospects and would try to understand their architecture for sending and processing webhooks. Webhooks had the perception of being an easy problem.

Many of them had gaping security holes, their architecture would fail to scale given average business growth over a few years, and they needed all the bells and whistles that make webhooks reliable. But yet, webhooks were still perceived as "easy". Being perceived as "easy" makes the sales job significantly more complicated.

It is comparatively so much better when the prospect thinks it's too difficult to do this and that they should look for external help.

Truths about discovered ICP

After many hits and trials with different customer segments, I narrowed down my focus on targeting software companies that are API-first products and have a publicly documented need to send webhooks typically in the seed, series A/B stage. This was the perfect sweet spot. Three prominent categories emerged:

  1. Ones who cared about the software they wrote: they didn't need my help, they would implement the webhooks features as they needed it and built it well enough to make sure it scales
  2. Ones who didn't care about the software they wrote: they did need my help, but they are rarely used to taking action, and since they don't care, the sales cycles were going to be very long
  3. Ones who were really early, had a very small team and often didn't have a lot of money. They often are validating whether their customers are going to be dependent on webhooks or not.

To effectively execute this idea, I would have to focus on the engineers and managers/VPs/CTOs who didn't care about the software they wrote, and internalizing this is how I would spend the next decade of my life scared me or spend a lot of my time and energy on free users who might churn and then rely on their growth and hope they eventually pay.

IMO, Svix started with the third category.

Distribution Strategy conundrum

I learned a lot of things about SEO and prospecting while working on Nirah. For SEO, some keywords are related to your business, but more importantly, the search intent matters, i.e., what happens in a person's life that they feel the need to search for. All the content writing and getting backlinks happens after that. Among many conversations, only a few people mentioned that they google how to build a webhooks system, which was evident since search volumes for webhooks-related keywords were quite low. People often search Google for specific company webhooks and usually go to their docs when consuming webhooks. Hookdeck, which helps in processing webhooks, understands the search intent very well and manages to get backlinks from the many companies' webhooks docs.

Prospecting works better as a distribution channel for this idea, although not as good as open source. I didn't prefer to go wide at the start. Instead, I choose to be hyper-specific as to whom I will serve. Top-down only works when the problem I am solving is one of their top 5 things to do, and after many calls, webhooks didn't make it to the top issues they care about. Bottom-up works best when coupled with open source. So what I ended up trying was "middle-out". I didn't want to build yet another open-source webhooks product.

Svix and Convoy got more adoption than all the previous webhook startups because of the open source aspect which gave them plenty of eyeballs.

Lack of effective positioning

Ask any founder or VC, and they would love to tell you about the importance of positioning. It is crucial to build and structure products and distribute them around what people like and how they want to buy them. Because I had picked an easy problem to work on, the surface area of the things I could do around positioning was quite limited, which meant picking something without having many options. It's not a great situation to be in.

Svix and Convoy (two of my competitors) have an open-source product with similar GitHub stars and have effectively identical positioning. I knew one was better than the other, yet I saw people failing to recognize which was better and choosing the one that was not reliable. This concerned me since I was building a closed-source enterprise product, and the prospects would only have to rely on social signals to decide what to buy.

Long-Term Competitive Advantage

A long enduring competitive advantage that lasts for multiple decades is a truly rare beast. Having "better tech" is not a long-term competitive advantage since if the problem is of high value, one would always face competition, and tech can be copied or made even better.

Some significant durable competitive moats a company can have:

  1. Network effects (Mastercard/Visa, social networks)
  2. Economies of scale and scope (Amazon, Walmart, D-Mart, Big Cloud)
  3. Switching costs (Microsoft, Big Cloud)
  4. Last mile coverage (Asian Paints, Hindustan Lever)
  5. Platform (Windows, Mac, iOS, Android, Linux)
  6. Intellectual Property until they expire

Unfortunately, this product had no scope to build a genuinely durable competitive moat. I could have gotten a start, had some cashflows, which would have sucked up years out of my life, and then focused on something with scope to build a long-term competitive advantage, but it is so much better to start with that.

Free & Open Source

A surprising number of prospects asked me whether the product I was building was free and open source, and to their disappointment, I said no, and for many, it appeared to me that I had lost their interest. I took it as a signal that webhooks are not something they would pay for, and I kept looking for other industries and other use cases. Having a pricing page and a clear idea not to pursue open source as a distribution strategy helped me be upfront with the prospects and understand their willingness to pay and the importance of the problem.

In speaking with a friend, he mentioned that this is common with Indian customers.

Surprisingly, when I was working on Reverse Penny Drop - RPD for Setu, nobody asked if the product was going to be free or open source, and most of them paid the first price we quoted them. These were all Indian customers.

A misconception I had about myself

When I started my entrepreneurial journey, I wasn't hell-bent on picking a huge market or a big hairy problem, and I thought I would be comfortable if the market turned out to be smaller than I had imagined. During the journey, I realized that I would never be content spending an entire decade trying to build a lifestyle business.

There's nothing right or wrong about running a slow, boring business. There are many times I would still prefer to invest in a slow, boring business than chase the fastest-growing thing.

But if I had to run the lifestyle business, it would become a drag for me.

What do breakout B2B SaaS products look like?

The best B2B product gets the buyer, user, or both promoted at their organization for adopting/buying your product. The core reason is that the product they purchase should help them with the business metrics for which they are responsible. So when they buy the product and their business metrics improve, they look like a genius at their org.

It's very evident in the first call to everyone why it benefits everyone involved. The other calls are required to validate whether it works, price it, convince internal stakeholders, provide support, etc.

Because of my involvement in RPD, I had a front-row seat to see what breakout products look like and how they are molded. RPD was focused on stock brokers, where Zerodha was the top discount broker, and Groww, Upstox, Dhan, PayTM money, etc, competed against them. Their internal business metrics were heavily focused on customer acquisition. RPD helped them acquire customers faster, which we did prove to them by assisting them in measuring how much it helped them (creating documentation around it helps improve retention since new people join organizations and they don't know the context).

After a year of buying RPD, Groww crossed Zerodha in customer acquisitions, and it was a celebration for their team. They would have done many things to ensure this, and not all credit goes to RPD, but it helped them. A lot of people at a lot of companies got promoted because of RPD.

What would I do differently next time?

Pick a hard problem that is extremely useful to the world. If the world ever needs those services, I should be the only game in town.

My managerial style

This was the first time I had a direct report (founding engineer). My managerial style can be summarised as:

  1. Be in a position to pick (hardest and most important)
  2. Get good at picking
  3. Hands-off by default (except for onboarding), hands-on whenever essential
  4. Be direct, ask others to be direct
  5. Make sure people share bad news with you as it happens; good news can wait, but bad news can't wait
  6. Help improve the personal growth rate of people


Selling ITC stake

The ITC stake had more than doubled in the short time I have held it. I remember I purchased this in the 21-22 bull market, and I was somewhat disappointed that in the biggest bull market in history, the only idea I could find was buying ITC when all VCs were funding companies left and right at crazy prices. Usually, the most conservative ideas in the biggest bull markets are the best-performing ones. I sold the ITC stake to fund my startup, Nirah Webhooks.

Selling JM Financial

JM Financial had gotten into real estate lending heavily after 2013. Till 2020, they had never shown NPAs. When the COVID wave hit, the management team was confident that their loan book would remain unscathed, and instead, they would lend more to buy out other players. However, as the moratorium lifted, it was evident that their loan book had gone bad, and their NPAs had spiked to 6%. Their Chennai real estate developers' book showed 25% NPAs. They immediately put the real estate story in the back seat without explaining what happened to their shareholders and started promoting their wealth management story. They had done some business restructuring, which was to obfuscate things. I wouldn't say I liked this lack of accountability, and I no longer felt they deserved my capital, so I sold.

Adding to IDFC First Bank

IDFC First Bank gave 5-year guidance when the merger between Capital First and IDFC Bank happened in 2018. They followed through with the guidance and had met them. They had done everything right, taking provisions whenever required and investing in the future by hiring great people and building the initial systems and bank branches in those five years. Their numbers reflected the story they told investors. I liked this behavior, and when they came up with new guidance for the next five years, I was inclined to believe them. I felt they deserved my capital, so I added to my position. It is now the most significant position in my portfolio.

What's next?

Heading into the coming year with some experiments and personal projects I would like to pursue. I wonder if those experiments/projects would turn into a startup and whether I will take up a job. I am going to write more on my website this year. Although I have been posting annual reports regularly, the blog section has taken a back seat for many years, but not anymore.