Learnings from first 6 months of starting up

I was trying to solve for size related returns and conversion drop off for online ecommerce from Jan-June 2024. I stopped working on it recently. If someone wants to use the product you can reach out or play with it for free here. Here are some things I have learnt.
The only way to learn from some mistakes is to make them?
I made a lot of common first time founder mistakes, including the ones I was explicitly told not to make. I decided there was a problem without talking to people and started working on it. Since it was B2B2C, I was convinced it was a consumer problem as I and people around me faced this problem. I roughly knew the numbers of returns in online apparel (30%) and that incorrect size (60% of 30%) was the biggest factor so I concluded that it must be a business problem too. The customer calls I had were just me trying to get these preconceived ideas validated. I jumped to building what I thought was the right solution - an image based size predictor. Cooked up theories on why 2024 is the right time for such a solution from tech perspective. Talked to many leading 3D Human Shape Estimation researchers to get in deep details of how to solve this. All of this only to realize the image based solution is a bad one as the effort spent for the marginal increase in accuracy was low. Note to future self: never write code before someone agrees to pay for what you’re building. Unless, you’re building for fun in which case you do you king.
 
💡
You need to find a real problem to create a business. You need to find a real AND burning problem to create a venture scale company.
 
The best salespeople aren’t great at selling whatever you’d give them. They’re the best at finding out what sells easily.
I didn’t fully understand “sell before you build”. Most brands used to tell me to show the accuracy first and then they’ll use my product. Once I spent 45 days building it, some of those brands never even took the demo call. I now realize how important it is to have 1-3 people ready to buy your solution even before you have it. It shows that the problem is big enough for someone that they’re ready to commit to just the idea of someone solving the problem (fka hair on fire problem). This also helps build the right version of the product. I think finding these people (fka design partners) is much harder than building the product so there’s temptation of skipping it.
Selling is hard. I realised how important is to narrow down who is your customer. This means finding which out of different categories, geographies and size of brands would be the most interested. This also means who in that brand would be the right person to reach out to. Someone wise said, “For B2B, build something that will get someone in your target company promoted”. Selling to product manager was not the best thing as they lack the decision making power and have so many other tasks for them to priortize this. Largely reaching out to CXOs or Director/VPs in big companies is the way to go. If they pass the lead to someone below them it could be a sign it’s not important enough for them to look at. I also saw first hand that sales is a numbers game and you should not take rejections personally. I assumed if I can’t convert even one of 20 customers, people probably don’t need it. I should have reached out to everyone I could. This was hard honestly as I think I was afraid of more rejections. Sales is a numbers game. I am yet to fully interalize it.
 
💡
If I were to do B2B SaaS again, I’d just find a problem where I can find at least 2 companies willing to pay a decent money even before there’s a product. Even better if they are willing to work together to build this. If 0/100 companies are ready to be design partners, it’s probably not a burning problem.
 
Get a good sense of the space: talk to everyone who has succeed or failed doing what you’re trying to do. Talk to VCs to get numbers. Talk to founders of adjacent companies in the same space to find challenges and share notes. I realised that Ecomm SaaS is not a big space 4 months into this after speaking to 5 founders who were building different products as microSaaS. Only companies providing large values to their customers like Shopify and Klaviyo have become venture outcome in this space. If you’re selling something that adds X value to your customer, and there’s some other company doing something which provides aX value to the same customers, talking to the founders of this company can tell a lot of things. This is not to say never do Ecommerce SaaS, but it’s harder to raise money here which makes it harder to play the venture capital game. Only do it if you have really high conviction.
Find people to work with: The general problems with solo founders “you need someone to discuss stuff with” and “morale” can be managed as a solo founder[2]. However, it’s hard for an inexperienced solo founder working in a new industry to move fast. It’s also one of the privilge of starting up to choose and work with people you like and admire. Fwiw, it’s not horrible to start things alone[3]. Good cofounder > Solo > Bad cofounder.
 
Specifically, talking about findmysize:
TLDR: I still think it’s a good problem to work on as a MicroSaaS solution that can cross $1Mn in ~3 years (when coupled with couple of small adjacent products[1])
The idea was that returns is a big problem in online apparel space - 30% of clothes on average are returned. On paper, about ~60% of them happen because of size related issues. This number is potentially inflated by 10-30% because of the perceived ease of choosing size as the reason.
Only a few people actually bother to look at size charts and even if they did, not many people confidently know their body measurements. The way most users shop is that they have some sense of their size i.e. they think size “M” fits them generally. Irrespective of where they’re shopping from, they will just buy “M” and hope it’s right. They don’t mind the process of returning so much. People who don’t like returns generally prefer offline shopping. Some people order 2 sizes that is M+L for each piece and then return the ones they don’t like.
The market didn’t feel big enough
Even globally, the market is not big enough for a venture outcome. One way to gauge this is so far the biggest company (True Fit) solving this is at $10Mn ARR and no significant change has happened/forecasted which should change this. Even though this is not necessarily a reason to stop working on something, it’s an indicator that it would be hard to raise money. Adjacent solution like virtual tryons have seen a similar fate so far. In a good case scenario, such a solution would have increased conversion by 20%. They’re several other companies doing all kinds of things to give the same value. I have spoken to them and they’re not trying to play the venture game at all because you can only charge your customers $99/month and not everyone is interested to buy it even then.
The problem is not urgent and not deep enough to expect high growth
If you could tell people the right size when they’re ordering you could add value in two ways:
  1. People would buy more online i.e. increase in conversion
  1. There would be less returns which saves money for the company and makes for a better user experience
For the customer of FMS which are apparel brands, 1 is a significantly stronger incentive than 2. 1 would directly increase conversion and topline. As mentioned above, increasing conversion by 5-20% is, to my surprise, not a lot of value for brands. It, therefore, doesn’t guarantee a venture outcome in this space. It’s also hard to attribute conversion. Plus this in an extra step for the user which makes brand worry about drop off and not worth testing also. Talking about 2: Venture funded brands or startups in general care a lot more about revenue (GMV) than profits for the first many years. Surprisingly, any order that is returned is still considered a part of GMV. Technically, if you reduce your 30% returns by 50% next quarter, you would have 15% less GMV. If you’re solving for growth in GMV, you wouldn’t necessarily want that. This is still difficult to digest for me. Only more mature brands who shift focus to sustainability and margins seem to have reducing returns as a priority. Brands are a lot more interested in solving for AOV and conversion. What another Ecomm founder told me recently is that I should have gone after legacy brands like Arvind Fashion, Aditya Birla[4], etc.
On speaking with the head at TMRW house of brands, I found out they’re only interested to try this for Bewakoof which is the more mature brand. Younger brands, they said, had many simple problems like “long delivery time”, “inconsistent manufacturing”, “suboptimal user experience”, etc. Which were bigger priorities than to try something novel. I assumed younger brands might try this to stand out or come across as more credible, but the counterarguement is they want to hide themselves in the traditional flow so as to not come across as unusual. They like to follow bigger brands.
If at all worth trying, US is much better
India, as many people recommended, is not the ideal place because:
  1. limited culture of using Shopify apps
  1. avoid paying as much as possible
  1. have tech talent to build themselves or at least think they can build themselves
Even if it was harder, I should have made the effort of trying to US brands to begin with.
 
References:
[1] One problem we stumbled upon is most brands have low adoption for size charts. This could go as low as 2% for marketplaces like Myntra. So Myntra now shows the important measurements for each size when you hover or click on the individual size so the user is bound to see the measurements before ordering. We built this solution for Shopify brands. This is complimentary to FindMySIze and can be sold together.
[2] This might seem like a general life advice and maybe it is, but building a support system as a solo founder can mitigate the morale issues. Other than having a support system, it can be helpful to have a personal board of directors. Find people, ideally from startup ecosystem, who are smart and rational who can ask questions and point to things you might miss. Send them a weekly update on what you have done and what you plan to do. Ask them what you might be doing wrong. This not only adds constraints which make you accountable, but also make you write (i.e. think) deeply on a weekly basis.
[3] There are some benefits I have noticed of starting up solo initially. I built a lot of confidence doing everything end to end. Likely I would have said, “I can’t do tech” if I didn’t spend some time building the whole product myself. I ended up meeting a lot of people in search of a cofounder which I hope in the long term is going to make it easier for me to find people to work with in different capacities.
[4] ABFRL is using a solution like this - Prime AI which is a UK based company.
I was trying to solve size-related returns and conversion drop-offs for online e-commerce from January to June 2024. I stopped working on it recently. If someone wants to use the product you can reach out or play with it for free here. Here are some things I have learned.
The only way to learn from some mistakes is to make them?
I made many common first-time founder mistakes, including the ones I was explicitly told not to make. I decided there was a problem without talking to people and started working on it. Since it was B2B2C, I was convinced it was a consumer problem as I and the people around me faced this problem. I roughly knew the numbers of returns in online apparel (30%) and that incorrect size (60% of 30%) was the biggest factor so I concluded that it must be a business problem too. The customer calls I had was just me trying to get these preconceived ideas validated. I jumped to building what I thought was the right solution - an image-based size predictor. Cooked up theories on why 2024 is the right time for such a solution from a tech perspective. Talked to many leading 3D Human Shape Estimation researchers to get into deep details of how to solve this. All of this only to realize the image-based solution is a bad one as the effort spent for the marginal increase in accuracy was low. Note to future self: never write code before someone agrees to pay for what you’re building. Unless you’re building for fun in which case you do you, king.
<aside> 💡 You need to find a real problem to create a business. You need to find a real AND burning problem to create a venture-scale company.
</aside>
The best salespeople aren’t great at selling whatever you give them. They’re the best at finding out what sells easily.
I didn’t fully understand “sell before you build”. Most brands used to tell me to show the accuracy first and then they’ll use my product. Once I spent 45 days building it, some of those brands never even took the demo call. I now realize how important it is to have 1-3 people ready to buy your solution even before you have it. It shows that the problem is big enough for someone that they’re ready to commit to just the idea of someone solving the problem (FKA hair on fire problem). This also helps build the right version of the product. I think finding these people (FKA design partners) is much harder than building the product so there’s the temptation of skipping it.
Selling is hard. I realized how important is to narrow down who is your customer. This means finding which out of different categories, geographies, and sizes of brands would be the most interesting. This also means who in that brand would be the right person to reach out to. Someone wise said, “For B2B, build something that will get someone in your target company promoted”. Selling to the product manager was not the best thing as they lacked decision-making power and had so many other tasks for them to prioritize. Largely reaching out to CXOs or Directors/VPs in big companies is the way to go. If they pass the lead to someone below them it could be a sign it’s not important enough for them to look at. I also saw firsthand that sales is a numbers game and you should not take rejections personally. I assumed if I can’t convert even one of 20 customers, people probably don’t need it. I should have reached out to everyone I could. This was hard honestly as I think I was afraid of more rejections. Sales is a numbers game. I am yet to fully internalize it.
<aside> 💡 If I were to do B2B SaaS again, I’d just find a problem where I can find at least 2 companies willing to pay decent money even before there’s a product. Even better if they are willing to work together to build this. If 0/100 companies are ready to be design partners, it’s probably not a burning problem.
</aside>
Get a good sense of the space: talk to everyone who has succeeded or failed in doing what you’re trying to do. Talk to VCs to get numbers. Talk to founders of adjacent companies in the same space to find challenges and share notes. I realized that Ecomm SaaS is not a big space 4 months into this after speaking to 5 founders who were building different products as microSaaS. Only companies providing large value to their customers like Shopify and Klaviyo have become venture outcomes in this space. If you’re selling something that adds X value to your customer, and there’s some other company doing something that provides aX value to the same customers, talking to the founders of this company can tell a lot of things. This is not to say never do Ecommerce SaaS, but it’s harder to raise money here which makes it harder to play the venture capital game. Only do it if you have a really high conviction.
Find people to work with: The general problems with solo founders are “you need someone to discuss stuff with” and “morale” can be managed as a solo founder[2]. However, it’s hard for an inexperienced solo founder working in a new industry to move fast. It’s also one of the privileges of starting up to choose and work with people you like and admire. Fwiw, it’s not horrible to start things alone[3]. Good cofounder > Solo > Bad cofounder.
Specifically, talking about findmysize:
TLDR: I still think it’s a good problem to work on as a MicroSaaS solution that can cross $1Mn in ~3 years (when coupled with a couple of small adjacent products[1])
The idea was that returns are a big problem in the online apparel space - 30% of clothes on average are returned. On paper, about ~60% of them happen because of size-related issues. This number is potentially inflated by 10-30% because of the perceived ease of choosing size as the reason.
Only a few people actually bother to look at size charts and even if they did, not many people confidently know their body measurements. The way most users shop is that they have some sense of their size i.e. they think size “M” fits them generally. Irrespective of where they’re shopping from, they will just buy “M” and hope it’s right. They don’t mind the process of returning so much. People who don’t like returns generally prefer offline shopping. Some people order 2 sizes that are M+L for each piece and then return the ones they don’t like.
The market didn’t feel big enough
Even globally, the market is not big enough for a venture outcome. One way to gauge this is so far the biggest company (True Fit) solving this is at $10Mn ARR and no significant change has happened/forecasted which should change this. Even though this is not necessarily a reason to stop working on something, it’s an indicator that it would be hard to raise money. Adjacent solutions like virtual tryons have seen a similar fate so far. In a good case scenario, such a solution would have increased conversion by 20%. They’re several other companies doing all kinds of things to give the same value. I have spoken to them and they’re not trying to play the venture game at all because you can only charge your customers $99/month and not everyone is interested in buying it even then.
The problem is not urgent and not deep enough to expect high growth
If you could tell people the right size when they’re ordering you could add value in two ways:
  1. People would buy more online i.e. increase in conversion
  1. There would be less returns which saves money for the company and makes for a better user experience
For the customers of FMS which are apparel brands, 1 is a significantly stronger incentive than 2. 1 would directly increase conversion and topline. As mentioned above, increasing conversion by 5-20% is, to my surprise, not a lot of value for brands. It, therefore, doesn’t guarantee a venture outcome in this space. It’s also hard to attribute conversion. Plus this in an extra step for the user which makes the brand worry about drop-off and not worth testing also. Talking about 2: Venture funded brands or startups in general care a lot more about revenue (GMV) than profits for the first many years. Surprisingly, any order that is returned is still considered a part of GMV. Technically, if you reduce your 30% returns by 50% next quarter, you would have 15% less GMV. If you’re solving for growth in GMV, you wouldn’t necessarily want that. This is still difficult to digest for me. Only more mature brands who shift focus to sustainability and margins seem to have reducing returns as a priority. Brands are a lot more interested in solving for AOV and conversion. What another Ecomm founder told me recently is that I should have gone after legacy brands like Arvind Fashion, Aditya Birla[4], etc.
On speaking with the head at TMRW House of Brands, I found out they’re only interested in trying this for Bewakoof which is the more mature brand. Younger brands, they said, had many simple problems like “long delivery time”, “inconsistent manufacturing”, “suboptimal user experience”, etc. Which were bigger priorities than trying something novel. I assumed younger brands might try this to stand out or come across as more credible, but the counterargument is they want to hide themselves in the traditional flow so as to not come across as unusual. They like to follow bigger brands.
If at all worth trying, the US is much better
India, as many people recommended, is not the ideal place because:
  1. limited culture of using Shopify apps
  1. avoid paying as much as possible
  1. have tech talent to build themselves or at least think they can build themselves
Even if it was harder, I should have made the effort of trying to US brands to begin with.
References:
[1] One problem we stumbled upon is most brands have low adoption for size charts. This could go as low as 2% for marketplaces like Myntra. So Myntra now shows the important measurements for each size when you hover or click on the individual size so the user is bound to see the measurements before ordering. We built this solution for Shopify brands. This is complimentary to FindMySIze and can be sold together.
[2] This might seem like general life advice and maybe it is, but building a support system as a solo founder can mitigate the morale issues. Other than having a support system, it can be helpful to have a personal board of directors. Find people, ideally from the startup ecosystem, who are smart and rational and can ask questions and point to things you might miss. Send them a weekly update on what you have done and what you plan to do. Ask them what you might be doing wrong. This not only adds constraints that make you accountable but also makes you write (i.e. think) deeply on a weekly basis.
[3] There are some benefits I have noticed of starting up solo initially. I built a lot of confidence doing everything end to end. Likely I would have said, “I can’t do tech” if I didn’t spend some time building the whole product myself. I ended up meeting a lot of people in search of a cofounder which I hope in the long term is going to make it easier for me to find people to work with in different capacities.
[4] ABFRL is using a solution like this - Prime AI which is a UK-based company.