At Episode 1 we think of ourselves as stage specialists, investing in amazing founders with compelling insight into important problems. We try to be the best partner in leading them on the journey from Seed to Series A, and we invest opportunistically with no overarching thesis.
However, we do research ideas we like and develop patterns we see in the market, and we wanted to share some of our thoughts in this multi-post blog.
Theme 3: Smart markets
During the first great Internet investment cycle – yes, the one that ended in the infamous dotcom crash – I was working for a VC company in the Bay Area that had the bright idea to focus away from the consumer space and onto B2B problems. Their idea was to systematically look at industrial markets and find where inefficiency was the greatest, insert marketplaces and make a percentage of every transaction.
The markets for dental supplies, or office equipment, or mining drills were large and to our eyes inefficient – there was lots of suppliers, non-transparent pricing, discovery was hard, transaction fees high. The Internet was the obvious solution to transform them all – and we were going to sit in the middle making money. Brilliant! We even hired fancy consultants to systematically go through supply chain after supply chain to find the best targets. We funded lots of companies and business plans. And for a while the idea had huge currency – a bit like the “sharing economy” or the infinite number of taxi hailing apps has now – with lots of strategic investors wanting to invest in our fund.
15 or more years later there are very few of these companies left – and with the benefit of hindsight there are lots of reasons why in fact the basic idea didn’t work very well. People have captive suppliers for historical reasons. Agents act as market makers, absorbing surplus inventory or pricing shocks. Inventory is not mobile – e.g. supply chains are locally integrated. There are branding or other issues around the disposal of inventory – it has our name on it, we would rather destroy it than have it appear in the grey market. Perhaps most importantly markets tend to cluster around the major power nodes – the largest buyers or sellers. And they typically do not want to allow anyone in to control the market.
“Markets tend to cluster around the major power nodes – the largest buyers or sellers. And they typically do not want to allow anyone in to control the market.”
Now that might be a long introduction to this theme – but it is to make the basic point quite clearly that there has to be a very good reason as to why a marketplace is the right answer to your problem. Because you can technically do something doesn’t in anyway mean that you can operationalize it – and there may well be very good business reasons why no market has existed here before. And these are in turn quite unlikely to be addressed by your shiny new app.
So the basic point is a new business will only solve a market problem where there is something unique that only the latest technology can do. And the good news is there are plenty of things that fit that bill. There are still large information inefficiencies such as discovery or pricing in many markets, and you have all that wonderful mobile immediacy and geo-location data to support you. We’ve seen a number of interesting B2B applications in shipping or logistics recently – at any given point there are plenty of empty trucks or ships that could take your workload, but you don’t know where they are, whether they can reliably fulfill your service level requirements or if they even meet your financial parameters.
“A new business will only solve a market problem where there is something unique that only the latest technology can do.”
There is also a lot more contextual data and machine learning capability now that exists to help match supplier and buyer in more intelligent – smarter – ways. Perhaps you are looking for a dental technician for a few days. In North London, and who speaks Polish. Oh, and who has all the latest health and safety training and has a charming bedside manner. Or a back end developer interested in working on an agricultural application. And who perhaps grew up on a farm – and is maybe interested in moving back to one. Or a refrigerated truck with capacity for two containers to go from Manchester to Istanbul leaving tonight and arriving within 48 hrs.
The amount of data about participants and commodities in a market you can manage, the scale of the data sources and the powerful matching tools that are becoming available enable liquidity in markets like these that couldn’t exist before – the problem was too hard to solve, inventory highly perishable and discovery too hard. Fantastic UI, workflow that is geographically enabled, field based computing platforms, and substantial fragmentation on both sides of the market with identification through self service and verification and credibility established by participation in the market.
Obviously the first question any ambitious start-up needs to answer having identified an interesting marketplace is why has this not been automated historically? And assuming the answer is it was too hard until the right technology came along, the next question is how is your technology solution going to solve this? Indeed, we think there is a fair amount of technology in these kinds of plays, so we like to see a strong tech team. We are also likely to want to know what you are doing with all the data you are collecting, who owns it, how secure and regulated it is – and of course how useful would other participants in the market find analytics derived from it?
“the first question…is why has this not been automated historically?…the next question is how is your technology solution going to solve this?”
Ultimately though it is the nature of the market that matters – and your ability to access it. Who has pricing power? Who are the anchors? Are the providers of goods or services truly independent? Are there geographic, regulatory or standards attributes that would impact liquidity? We want markets where the largest buyers and sellers are labeled “other” and where there is a substantial degree of standardization or at least description of the standard trading commodity – a description is machine interpretable. And maybe then you really can sit in the middle and count the money.