Friday 26 April 2019

Startup Diary: Billion-dollar question: do the maths on your market before you pitch up

Stock image
Stock image

Richard Rodger, founder voxgig

There's a time for every startup to pitch to investors, and our time is now. We're raising a seed round, and I'm putting together our pitch deck. As this is a real-time startup diary, you get to follow along as I build the slides. We're only one slide down so far - the business model, so perhaps we should pick up the pace? No. Before we cover the classic slides (problem, team, competitors, etc), we need to lay some foundations first.

Your deck should consist of two sections. The first section is the pitch. This is what you show to investors as standard. You're looking at 20 or so slides maximum. Many guides to pitch decks will give you exact and specific lists of slides, with as few as 10 slides in some cases.

I think you have to build your own deck and put together a pitch that you can give credibly. Assimilate all the advice and tune it to your own needs.

The second section is an appendix, where you place more detailed slides that are text or number heavy. These can be used when an investor wants to 'go deep' on a topic. These are the best slides to start with, as your primary pitch slides should tell a story based on the details covered in the appendix. This approach also means that you know you've always got evidence to back up your assertions, and can pull up a relevant appendix slide if necessary. That helps very much with the confidence that you project in your pitch.

There is one awful question that you must be able to answer in less than 100 milliseconds: what is your TAM? The Total Addressable Market (TAM) of your startup idea is one of the big sense-checking numbers that investors use to make a decision. There is no point investing in a great idea with a great team and completed product, if the total sales of that product are too small to generate 'venture returns'. You can make a good living from running a restaurant or a pub, or finding a unique angle on a boutique clothes or furniture shop, but you're never going to generate 'venture returns' - a hundred times the money put in.

The Total Addressable Market idea is something you should take seriously for your own sake - why put so much time and effort into a startup that can never be a billion-dollar company? It's a mistake near to my heart because I made it my first business.

After the dotcom boom and bust, and having worked as a junior developer in a few startups, I got the itch to start my own thing. In 2003 I launched an online business selling high-volume data-processing software components. I made so many mistakes that I could fill this weekly diary for a year writing about them. The very description of the business should already have made you say "oh dear".

Well, let me tell you, there was actually a business in that game - processing large volumes of comma-separated text files is something many companies need to do. But it wasn't a business that would ever scale to anything more than a so-called 'lifestyle' business. A quick calculation of the TAM would have told me that. Let's go back in time, and use my first little business as a case study.

My product was a downloadable software component that developers could use in their own projects to handle CSV files (Comma Separated Values). This is a common format for exporting data from databases, particularly older ones (you can create a CSV file yourself by choosing that as a export format from Excel). I used per-developer licensing. That means you paid about $100 per developer on your team. You could buy bulk licences that applied to your entire organisation. I was selling into an enterprise market.

As an aside, my pricing was utterly nuts! Think about it - your best customers are big companies, and you're giving them all the discounts. Big companies, when they really need something, are completely price insensitive. I could have made a lot more money increasing my prices x10. In fact, when I started out, my prices were so low I had a huge credibility problem because I looked like a 'one guy in his bedroom' operation (well, that's exactly what I was).

If we calculate the TAM, we can see that there's lifestyle business here, but not a venture business. To calculate the TAM, you use two methods: bottom-up, and top-down. These give you lower and upper bounds on the TAM. You can also benchmark your estimated range against similar large companies in comparable or adjacent markets to see if their annual sales are consistent with your workings.

For the top-down approach, you do a lot of 'secondary' research - industry reports, blog posts, journal articles, etc. You try to get your hands on as much numerical material as possible. In the case of the software component market of the early 2000's, Gartner estimated this market at about $7bn. That's ... not great. And it's not even a homogeneous market. You have to subdivide it by programming language and problem domain. Taking a rough cut, data-processing components for the Java language (which I used), probably had a TAM of $100m, being very generous. So that's our upper bound. It's already way too low for a venture-backed startup, by the way.

How do you calculate the lower bound? Try to estimate the total number of customers and multiply that by the price. This is analogous to a popular "brain-teaser" programming interview question: estimate the number of swimming pools in California. Substitute "swimming pools" and "California" to taste. It's supposed to show how you use deductive logic and estimation to make decisions in a low-information environment. It's so well-known now that the question is more useful in determining if the candidate reads the literature.

So how do you work it out? Start with gross estimation - I'm literally going to make this up as I write it. Population of California? It's a big state population-wise, and the US has at least 300 million people. Los Angeles has at least 10 million inhabitants, so let's say California as a whole has 30 million. How many have houses? First divide by two to handle cohabitation, then maybe 10pc have enough money to afford a house with a pool, and maybe only half of those actually have pools, so: 30m x 0.5 x 0.1 x 0.5 = 750 000 pools. Google tells me there are 1.18 million - I'm within an order of magnitude. And that's 'good enough'. This is about as accurate as you can get with bottom-up TAM calculations.

In the case of my first business, let's say at least the Fortune 1000 are interested, plus the half of the remaining businesses ranked by size, in the US alone - 500 000 companies? Of which say 10pc develop software in-house (this is the 2000s, not today's age of 'digital transformation'). Of which say 5pc have CSV processing needs, and of those maybe 75pc use Java (it was pretty popular). Total companies: 1875. These are still relatively big companies, and within those companies you'll have teams of developers. Say five developers a team, and 10 teams per company on average, and two of those 10 are working on data processing. Thus 1,875 x 2 x 5 = 18,750 potential licences. At $100 a pop, that gives a TAM of $1.875m. Thus we have a range of $2m-$100m (this range might seem wide, but it is more than sufficient to answer our ultimate question - is the TAM big enough for a VC?), and if we want to pick a point estimate, perhaps interpolate this to $20m (also adding the assumption that we could have 10X'd the prices). That's definitely a lifestyle business (you won't capture 100pc of that market - more like 1pc), and a decent one, but it's nowhere near what a venture capitalist wants to see.

I'm now going through a similar exercise with my current business, trying to understand the technology conference events space, and developing estimates using this approach. Obviously when you're building a real pitch deck, you have to put a bit more effort into your estimation models, but the basic principle remains the same.

One nuance you might also consider: what about the market you can actually reach? The so-called addressable market? Your global market might be $5bn, but if $1bn is in China, and another $1bn in India, you're probably not going to reach those markets before local competitors get there first. Also, the rest of the world doesn't just speak English, and it's very expensive to localise software, so the total 'serviceable' market might just be the Anglophone world. But none of this matters if your TAM is too small. Think big!

(Newsletter update: 3,854 subscribers, and an open rate of 12.5pc. So we hit a glitch in our subscriber invitation process last week - it's an opportunity to apply incident analysis to a non-technical part of our processes, which is a first for me. If you go back a few weeks to mid-August, I wrote a column about a technical outage that we had, and how we used that to improve our system. We'll use the same techniques here with our marketing processes - a topic to cover after we've pitch-decked ourselves numb).

Richard Rodger is the founder of voxgig. He is a former co-founder of Nearform, a Waterford-based technology consultancy

Indo Business

Also in Business