The case of the missing Canadian Dotcom crash
Or why we should all stop using Pitchbook for historical Canadian Tech Data
I’ve been thinking a lot about Venture Capital investment cycles recently.
Thanks for reading The Matt Roberts Newsletter! Subscribe for free to receive new posts and support my work.
I’ve started a new fund with my old friend David Dufresne and my (ex-Fasken Montreal) lawyer Gabrielle Paris. Our fund aims to raise over $50M to fill a gap for lead cheques in the seed stage. In other words, the goal is to lead or co-lead and, importantly, price every deal we do, working with our founders and co-investors to prepare our portfolio to raise their next round and execute their business.
The current venture market is weird, so I looked at how others have done it.
So, I’ve started looking at the backgrounds and history of venture funds in the more recent past. iNovia, Mantella, Golden, and VersionOne; all funds started in 2008-12, after the financial crisis. I then started looking at the (all be it thin) cohort from the post-dot-com period dating back to 2000, eight years before - Brightspark, Relay (where my colleague David worked), and others…
This led me down the rabbit hole of the similarities and differences between today's market and the previous downturns in our industry.
I’m not alone.
Recently, my former colleagues at BDC released a report entitled “Canada Venture Landscape” to seek “to understand how our ecosystem might respond to the current environment.”
Essentially reading my mind, the team at BDC:
looked at VC activity following past economic downturns, specifically after 2000 (dot-com era) and after 2008 (the Great Financial Crisis or GFC). We noted that after the dot-com bust, VC activity slumped from 2000 to 2004, at which point it had declined 51% from the 2000 high, recovering slightly in 2005.
Their commentary came with this helpful chart.
I’d like you to focus on the first part of the chart, the 2000-2010 period. Which is the period I’ve been looking at. I was surprised- by the numbers because they don’t correlate with what I’ve seen looking at the historical numbers. I think the chart is wrong because of one company.
Pitchbook was founded in 2007. It initially focused on Private Equity data in the United States, while its Canadian coverage focused on PE deal flow. It only started tracking Canadian VC data with any regularity in 2015, and its data locally only got acceptable to any degree in 2017 or so.
Pitchbook gets its data in Canada by scrapping Betakit, TheLogic, Globe & Mail, etc. It sometimes finds other less credible areas and inserts that into its Database. Sonder (which I led the seed round for in 2014) was my test case for how good their data was. It inevitably got the U.S. investors mostly correct but left out or ignored the Canadian participants. Pitchbook has been slowly adding data from the pre-2015 time, Mechanical Turk style, but that data pre-2015 is again only what it’s been able to find through Google searches.
And this is a problem.
Why? Well, newspapers and trade journals covered most of Canadian tech deals pre-2012. Today, those details are behind paywalls or at your local library… so they are essentially nonexistent if you are trying to find them today. Did it happen if it’s not online?
But since my pastime is Canadian tech history, I know where to look and it is online. Just not where Pitchbook looks; here in this is a government report on Venture Capital in Western Canada. It gives historical data for all Canadian Venture Capital going back to 1996 but using a different data set.
Notice any differences?
The dot-com (Year 2000) number for total VC invested in Canada here is listed as $6.4 Billion vs the $1.8 Billion in the Pitchbook-derived chart. If we use this data, Venture Capital collapsed in Canada from $6.4B in 2000 to $1.6B in 2003 or a decline of 75%; it recovered slightly after, until in 2009, it cratered again to less than $1 Billion.
To give a sense of how bad Picthbook’s data is, let’s look at my hometown when I started my career as an angel investor. Here is the Data for Ottawa in just Q4-2000.1
Innovance $115 M
Catena $90 M
Lantern Comm $89 M
Silicon Access $84 M
SS8 $38 M
Spacebridge $25 M
Quake Tech $18 M
Adherex Tech $10 M
Analog Design $8 M
Zucotto $53 M
Watchfire $38 M
Solidum Syst $25 M
Dragonwave $12 M
Internetivity $10 M
I got those deals from newspaper articles printed on paper. Those deals total $476M in one quarter (Q4) of 2000 in Ottawa alone, not including any other market. Using pitchbook data would mean that well over 35% of the total year 2000 Venture Capital in Canada was done in one city in just one quarter. And while Ottawa was the biggest market in Canada. This is just not believable. (for those asking, (all 2 of you) Ottawa received $1.23B in 2000.)
The second (older) chart uses data gathered on the ground and at the time. Thomson Reuters currently owns this data, but they don’t advertise it much anymore. They may own it, but it was put together by a firm named Macdonald and Associates, which was run by Mary Macdonald, who founded the firm in 1985 and ran it for 20 years. Thomson bought them in 2005.
Data methodology could account for some discrepancies - they often do. But it’s way too large, and I don’t think it does. Also, the CVCA used Macdonald & Associates data as recently as 2012 for its total market research2. If they had issues with their methodology, I think it would have been brought up then.
So, let us have some fun with Excel. These numbers I have from Macdonald & Associates, and where that ends (in 2016), I have the CVCA data going forward.
Then, I’ve added the data from the Picthbook numbers charted for comparison. I think the 2020-22 numbers are proprietary to BDC and not Pitchbook-derived. BDC would have much better data internally for what’s going on.
If you look closely, the Data of both charts started to merge in 2012 as Pitchbook improved and almost matches from 2015 onwards.
Accepting that the Pitchbook’s numbers are primarily correct from 2015 onwards and the second Thomson chart is correct before that date. We can say that Canadian Venture Capital peaked in 2000 and did not beat that number again until 2019. more importantly, we might also say that the Canadian Venture crash took over three years.
As an old guy, I find it hard to explain to today’s cohort of VCs what that period felt like to those on the ground in Canada. One way to start the conversation is to adjust the numbers for inflation, which gives a flavour of the fall from grace of Canadian Venture Capital. Here is the chart with each year adjusted to the 2023 equivalent dollar value.
This gives a better sense of how VCs of my vintage look at Candian Venture. When I first started in seed investments in 1996. We were running up to a peak. It has taken twenty years to get back to those numbers.
VC data, particularly the source of that VC data, is critical for us here in Canada.
Looking at the first pitchbook-generated chart could lead you to believe that while there was a decline, the domestic venture market was relatively small in the ’00s and only began to grow in the post-2010 period. This just isn’t the case. But not knowing that could lead you to draw very different conclusions as to how the venture market in Canada will react in this new reality post-2022, where we find ourselves now.
But all to say, don’t trust Pitchbook for your historical Canada Tech data.
In my next email, I’ll dive into what exactly drove that massive decline.
Ottawa grabs huge share of capital: Tech startups finished 2000 with staggering $1.23B invested: [Final Edition] Hill, Bert. The Ottawa Citizen; Ottawa, Ont. [Ottawa, Ont]. 13 Jan 2001: H1 / FRONT.