Things we learnt trying, and failing, to build an e-commerce platform for selling digital goods
Short version - It’s extremely hard to do
A year ago we set up Makeshift, a company dedicated to ‘giving a leg up to the little guy.’ Over the past year we’ve invented and built 11 products which have met with varying success. Over the next few weeks I’m going to try and write about the some of major things we’ve learnt by telling the story of these products, starting with our first app - Bitsy. If you want to get the next one, make sure to subscribe to our newsletter.
When we set up Makeshift we didn’t know what we were going to make. We had a bunch of ideas floating around, some of which were bits of ideas from previous projects, others which were blue sky ideas we’d always wanted to build.
One idea that Stef and I had been talking about for sometime was a concept to try and build the smallest digital thing we could think of that could make money—the idea was “sell a link”.
Sell a link
The conceptual purity of this idea was fantastic. A link. The unit of the internet. For sale. Sell. A. Link. Amazing. The product strategist in me just wanted to call time on the whole thing. Congratulations, you’ve found the end of the internet. Cheque please! But as you’ll see, it was, alas, not nearly as simple as we imagined.
We’d been really inspired by a service that had just launched in the states called Gumroad that let people sell any digital file using their own social media channels, and initially we wanted to take their idea and simplify it further—Gumroad hosted your file, and we thought we could refine that even further by just allowing you to sell access to a link.
What we did
So on the first day of the company we sat down together and discussed ideas for the product. We had a great name – Bitsy – and we started thinking about who it could be for. We had this grand plan which was that, even if Bitsy didn’t work initially, it was a good idea for us to build this payments solution first because then we could use it in lots of our other ideas that we had.
Above: Two TOP PRIORITY features
So we set to work – wireframes, customer development, prototype, tech spikes, sprints, the lot. At this stage we just had Jon working with us and a marketing specialist called Nathan.
We built a lovely little app
Within a couple of weeks we had a lovely little app that let you add a digital file from Dropbox or your computer, create a little landing page with a picture etc and send the customer a link after they added their email address (no payments yet). We asked our friends over for a demo.
The demo was fun. Our friends like us. We got some good questions. Mainly, great, but who will use this? So the next day we started thinking more about our marketing ideas – how would we grow this thing?
At first glance it seemed like a no-brainer – we’d growth hack the shit out of it right!? Start scraping lists of people who have digital things for sale on big sites and begin reaching out to them to explain the awesome simplicity of our product. We also experimented with Facebook ads, Google ads and a bunch of other paid growth ideas.
We also had this idea that we could also partner with site owners who could offer a ‘sell with Bitsy’ link to their users. But that turned out to be a massive black hole of feature requests - every site owner we spoke to had a million requirements, and if they had the chance to let their users sell stuff, they were already using an e-commerce plugin of some sort. Hmm.
At this point we started looking into ‘the payments bit’ and got a rude awakening. We’d seen that Gumroad was powered by Stripe, so we assumed, initially, that we could just do the same. The idea was that we would process the payments, hold on to the money for a bit and then repay the seller at the end of each month. So simple right! Wrong.
Anyway, we looked at Stripe. But Stripe wasn’t available in the UK or Europe at that point. Damn. So we looked at Paymill. But they were all over the place. So we phoned some bankers and some other payment processors to look at building our own solution. This is when we heard the dreaded term ‘third party payments aggregation’.
Third party payments aggregation. Hmmmmmmm…
It turned out that our idea, which seemed so simple at the start, was actually one of the most risky types of business for the payments processors, who are all very worried about fraud and chargebacks.
“It turned out that our idea, which seemed so simple at the start, was actually one of the most risky types of business around.”
The business category is called ‘third party payment aggregation’ (TPPA) and the typical example is ticket sales – if you sell tickets for someone else’s event, and they then cancel the event and don’t give you any money for refunds all the customers then charge back their credit cards to the bank, who then tries to pursue you.
For this reason TPPAs typically have to keep a large amount of money on deposit and have a very clear view of who the person they are selling on behalf of is - but in our case you could just sign up with an email address and start ‘selling’, and we had no idea of you had the rights to sell the thing you were offering.
Regardless, we carried on. If Gumroad could do it, surely we could. We got lawyers in. We contacted the CEOs of the various payments processors. Part way through Stripe came to the UK in beta and we immediately started hanging out with Andy and Diede - surely they would save us! Wrong. After much digging it turned out the problem went all the way to the very, very top - the difference between the regulatory frameworks for EU banks vs US banks and the degree to which you have to ‘know your customer’ in these markets.
Above: One of many annoying diagrams we had to draw for lawyers
So our great hope, Stripe, couldn’t offer the same deal to us in the EU as they did to Gumroad in the US. They were as annoyed as we were. But that didn’t help us.
We struggled through. We looked at setting up a US subsidiary. We started to have some good conversations with some more boutique providers, but we quickly realised that with them our margins were going to be much smaller than we’d planned for, because the extra risk meant the transaction fees were really high. The US thing without a business presence on US soil was a dead end due to international money laundering rules.
Selling stuff is hard work for users
Meanwhile, back on the shop floor we’d hacked in a dirty PayPal solution but we were having a lot of trouble getting people to sign up and actually sell something. We had a lovely site by this point, and the product pages were looking sexy. But who would sign up and use it?
Well, thinking mega vision we segmented our customer base, which of course was everyone in the world – sell a link, right! Right!! – into two big groups. People who definitely have something to sell (for example an eBook, an album, some design templates) and people who don’t yet realise they have something to sell (for example a spreadsheet, some coursework, some old design assets).
Lets take both these ‘segments’ in turn and explore why we couldn’t get traction within the tight timeframe we were operating against.
People who have something to sell
This seems like our obvious target customer. But its not a segment. As soon as you hit the streets (metaphorically) and start trying to acquire these customers you realise they don’t see themselves as ‘someone with something to sell’ but as a writer, a film maker, an artists, a musician, a designer … and each one of these groups has unique display and fulfilment needs around the specific thing they are selling, not to mention regional variations (one example - americans like to sell in dollars, Indians in rupees).
“People don’t see themselves as ‘someone with something to sell’ but as a writer, a film maker, an artists, a musician, a designer...”
Regardless of the unique features you need (which we don’t have) if you have something to sell, and you know it, you are probably already selling it. Which means switching to us is a ballache, and we have to compete on features and pricing. Did I mention we don’t have those features you need because we’re a universal service? Oh, and our pricing is really high because we can’t use Stripe. Crap.
Anyway, we broke the segment down by verticals and customer size and concluded that we couldn’t service the super stars, so we’d have to focus on the independents. But these guys presented another problem - they don’t care about selling stuff.
This was the final nail in the coffin for this segment. Because these independent small customers see themselves as musicians and artist they understandably want to spend their time writing music and creating art, not selling. So they are more than happy to give up a big chunk of their revenue in return for you finding the customer for them. This is why iTunes, Spotify et al are so big.
People who don’t have something to sell
So then we looked at the next, very intriguing segment – people who don’t realise they have something to sell. We thought of these people as people a lot like us, with hard drives overflowing with files that if only you could see some value in them you’d be happy to package up and sell. Boy were we wrong on this!
The first, massive problem with this idea is that the education cost is massive. You have to get people to understand the idea of Bitsy, then think about what they have to sell, then help them think about how to package it up. And then, if that’s not enough you’re asking them to go and find people to sell to - which leads to the next problem.
Above: A product from my Bitsy account. Pretty typical of our users - 313 views, 0 sales!
Even if you can do the above, which is really hard, you encounter a bit of an immovable problem - people who don’t realise they have something to sell don’t have social media accounts they can sell on and so no route to market. Think about it - if you have a twitter account or a Facebook page that you are normally posting selfies and holiday snaps on, its *really* weird if you suddenly start saying ‘hey, wanna buy my spreadsheet’. We had people asking if accounts had been hacked. It was awful.
Meanwhile, Gumroad were smashing it
All the way through this, we had another problem. We were trying to pretend it wasn’t true, but it was - we were basically copying a much more advanced, better funded company operating in a friendlier regulatory environment, Gumroad.
After a while it got to the point that we banned ourselves from visiting their site, because they’d solved every UI problem or feature idea so well it was really hard to not copy them, but when we did it was depressing. This wasn’t why we set the company up!
We failed, and we canned it
Anyway, it all came to head about three months after we started, and Stef, Paul and myself sat down in Albion and Paul said ‘I think we should kill Bitsy’. But the really interesting thing was that it wasn’t for any of the above mountain of reasons. Instead, he said ‘its because I’d rather make stuff that helps people create, not sell.’
This was an important moment for us. Stef and I had worked together for some time, so we had a good rapport, but we hadn’t worked with Paul very much, and, more to the point he was funding this whole thing. Back of my mind I was thinking, crap, we’ve already wasted loads of his money. So I said ‘but we’ve put so much money in already!’ He then said something else very important: ‘So what? I don’t think sunk costs should have any bearing whatsoever on the future.’
That was a big moment for me, and it laid the groundwork for what was to come - an explosion of new product ideas which we’d been brewing away whilst hacking on Bitsy. But that’s in the next post.
Above: Bitsy wireframes, in the bin with the rest of it
What did we learn that we took forward?
We learnt lots about payments, digital e-commerce and building apps for yourself, very fast. But a few key things I think we learnt that stayed with Makeshift after Bitsy were:
#1 - Don’t start with *only* a big idea in mind.
We were in love with the mega vision of ‘sell a link’ but this was an insanely big idea for the resources we had to command. We should have started with ‘sell an eBook in a very weird format that you can’t sell anywhere else’. Or something.
#2 - Don’t copy other people
This is more controversial. Being a fast follower works really well for some people. But not for us. At Makeshift we take pride in being makers, and that means we have to feel like we’re inventing new things everyday. If you are copying someone else, you just don’t get that buzz, and we need the buzz to motivate us every day.
#3 - Make something you can use yourself
This is a lesson that it took us a few more product iterations to learn fully, but with Bitsy we struggled to use it ourselves from the outset - and we were exactly the ‘don’t realise you have something to sell’ segment we were pinning a lot of our hopes on. We probably could have avoided wasting so much time if we’d realised this earlier.
#4 - Look under the hood of the business before building a product
This is sort of obvious. Of course you should do some market research before coding, but in this case there was a lot of complexity around payments and being a TPPA that we hadn’t anticipated. We could probably have spotted it earlier if we’d thought more about the business dynamics as much as the user experience at an earlier stage.
#5 - Stripe is an amazing business
This is kind of off topic, but doing lots of research and investigation into the various and incredibly painful ways online payments processing works Before Stripe (BS) meant that I got to realise just how disruptive Stripe is. I’m in awe of those guys. Big up yourselves. You’ve got one of the best businesses in the world.
Conclusion: Anyone want to buy Bitsy.io off us?
So, that’s Bitsy. All in all we spent about £65,000 on it over three months or so. Having said all the above, I think there’s a big opportunity to put a really simple front end on Stripe for people who only sell a few digital things. If you can hack the growth model.
So if you’d like to chat about buying the assets in Bitsy I’m all ears - we’ve got a lovely app, a nice brand (TM in US and EU) and a lot we can teach you. Ping me an email - nick at makeshift dot io.