How To Spot the Next Winning AI B2B Startup
Rosie from The Jetson’s is to blame for my early attraction to tech. And it’s Silicon Valley and the enterprise world’s fault for dragging me into a world that lives, breathes, and eats tech optimism. Yeah, I too believe in the idea of investing in, creating, and propagating the future through technology.
For 9 years, I pushed existing technology in the B2B Enterprise Tech space, at IBM, Cisco, and at Silicon Valley’s Glassbreakers. The thing is that as soon as I jumped into my last role as VP of Sales at Glassbreakers, I was introduced to this crazy group of people who, instead of pushing existing technology, were pushing for the technology of the future. Since then, I’ve been working towards becoming one of those crazies in the world of venture capital (it’s the - Steve Jobs - good kind of crazy).
Pivoting into this new road, I knew I needed to focus on a particular space, and I chose to put my eggs in the basket of AI — after all, it’s the one that I believe (with my body and soul) will bring the future to the world. And that’s what led way to a column on AI in the Huffington Post, an internship at K Fund, IE Business School and Kauffman Fellows Venture Capital (VC) programs, countless speaking engagements, and the opportunity to join Portfolia’s Enterprise Fund as an investor. And, I’m just getting started. Next (and post-MBA), I want to work for a VC fund full-time and invest in AI and emerging tech.
And so… as an aspiring VC associate in the AI B2B Enterprise space, I thought it would be interesting to analyze the AI B2B startup space. I asked myself, “what would be most valuable as a VC investing in AI?” My answer came down to: identifying features that make for a winning AI B2B Enterprise startup investment. And, right then, I set out to figure out how to spot the next winning AI B2B startup.
Methodology
Using reverse engineering thinking, it occurred to me that looking at AI B2B Enterprise startup exits may provide some insight. So, I went out to analyze all available AI B2B exits.
Exit Analysis
In this post I share my analysis of the AI B2B startup exits. I include insight on the analysis process, insight into the acquired startup, acquirer, investors, and perspective on the factors that may help you spot the next winning AI B2B startup. Going the extra mile, I share my predictions of next exits and share my perspective on what VCs need to keep in mind when looking for a high-potential AI B2B startup.
Analysis Process
I used Mattermark to identify all enterprise software startups that have exited in all history; keywords used included: artificial intelligence, machine learning, and deep learning. The algorithm spat out 31 AI B2B startups — 2 IPOed and 29 were acquired. Keywords in the Mattermark startup description included: analytics. I then used Crunchbase to find the founders of the startups and, finally, used LinkedIn to find founders’ educational backgrounds — degrees and schools they attended.
Exits
List of exited startups: Criteo, OpenSpan, Right Hemisphere, Proximal Labs, Koemei, Analyze Re, Cognitive Security, Metafor Software, Progress Software, QPID Health, Invincea, Acquisio, Cloudmeter, Wise.io, Kanjoya, Boomtrain, Algorithms.io, Nutonian, Prelert, Via Science, Brainspace, CyberFlow Analytics, SignalSense, Retail Optimization, ClearForest, Unified Inbox, Pattern, SkyData Systems, MetaMind, Cognea, and NudgeSpot (acquired by Boomtrain, which was also acquired).
More than a third focused on machine learning and analytics. The median number of founders was 2. The education of founders ranged from college to PhDs. Almost a third of co-founders had attended the same university. A large proportion of founder teams had computer science and engineering backgrounds. A third of startups had co-founders with MBAs.
The median time for an exit since startup launch was 5 years. The range of last funding before the exit ranged from seed to Series D. Post acquisition, over 75% of co-founders continued to work under the acquirer. Geographically, startups who exited were primarily headquartered in North America. 35% of them were from the Bay Area and over 15% were from Boston.
Money
Exit amounts were not widely available (although we do know that MetaMind was acquired for $32.8 Million and OpenSpan was acquired for $52.3 Million). The median of total funds raised by the startups before exits was $6.5 Million. Nearly half of the startups raised under $10 Million. OpenSpan, Invincea, and Criteo were the anomalies having raised $31 Million, $47.3 Million, and $63.4 Million, respectively. Before the startups exited, the last round of funding ranged from $100 Thousand (Koemei) to $40 Million (Criteo). The median amount of last funding raised before exits was $4 Million. Over 45% of startups fell in an annual revenue range of $5 Million to $10 Million at time of exit; and over 21% of them fell in a revenue range between $1 Million and $5 Million.
Acquirers
The acquirers list included: Jive Software, Lumendata, Salesforce, Boomtrain, SAP, Webroot, eviCore Healthcare, Splunk, GE, Crelogix, Zeta Global, DataRobot, Elastic, Workday, Sophos, Web.com, Ultimate Software, Pegasystems, Cisco, Reuters, Cyxtera Technologies, Revionics, and IBM. Splunk rose as top acquirer in the bunch with 3 startup acquisitions: SignalSense, Metafor Software, and Cloudmeter.
Investors
There was no pattern identified in the investor list; although, it was curious to see that Marc Benioff invested in MetaMind before Salesforce acquired the startup. Sequoia, In-Q-Tel, 500 startups, Matrix Partners, and Sierra Ventures, all invested in two of the startups analyzed that exited.
List of investors included: Allegro Venture Partners, Baseline Ventures, Constantin Partners, D.E. Shaw & Co., Floodgate Fund, Ron Conway, SV Angel, Tom McInerney, DFJ Gotham Ventures, Rally Ventures (formerly Icon Venture Partners), Sequoia Capital, Valhalla Partners,500 Startups, Munich Venture Partners, NVIDIA, Sequoia Capital, Sutter Hill Ventures, Khosla Ventures, Marc Benioff, Aeris Capital, Dell Ventures, Grotech Ventures, Harbert Management Corporation, New Atlantic Ventures, ORIX Venture Finance, Meakem Becker Venture Capital,Credo Ventures, Evolution Equity Partners, Cardinal Partners, Massachusetts General Physicians Organization (MGPO), Matrix Partners, New Leaf Venture Partners, Partners Innovation Fund, Ignition Partners,Tola Capital,Trilogy Equity Partners, GrowthWorks Capital,Vanedge Capital, Fonds de solidarite FTQ, Tandem Expansion Fund, Wellington Financial, AngelPad, Cota Capital, Jumpstart Ventures, Lerer Ventures, Signature Capital Securities, Streamlined Ventures, FTV Capital, Globespan Capital Partners, Imlay Investments, In-Q-Tel, Sigma Partners, Sigma Prime Ventures, Angel Investors LP, Plug and Play Ventures, Siemens Venture Capital, Toshiba America Electronic Components, Felicis Ventures, First Round Capital, SoftTech VC, Kae Capital, Alchemist Accelerator, Silicon Valley 2014, Voyager Capital, Atlas Venture, Fairhaven Capital Partners, Intel Capital, Sierra Ventures, Green Park & Golf Ventures (GPG Ventures), Medina Capital, MassChallenge, BDC Venture Capital, Innovacorp, Jevon MacDonald, Rho Canada Ventures, Connecticut Innovations, Elm Street Ventures, Adams Street Partners, Bessemer Venture Partners, Elaia Partners, IDInvest Partners, Index Ventures, Sapphire Ventures, and SoftBank Capital.
M&A Professionals: Predictions of Next Exits
If past trends continue, common features of startups that will exit next will include:
- Annual revenue $5–10 Million range
- Median of total funds raised $6.5
- Co-founders with engineering and computer science degrees and, at times, PhDs
- 1 to 2 founders
- Median of 5 years old
In order to identify potential AI B2B startups exits, I captured data from Mattermark, Crunchbase, and LinkedIn and analyzed it. I used Mattermark to identify all AI B2B startups that have not exited (keywords used included: artificial intelligence, machine learning, and deep learning). Mattermark outputted nearly 200 existing AI B2B startups. I used Crunchbase to identify founders and LinkedIn for educational backgrounds.
Predictions of Next Exits
If past trends continue, the startups that have 1 to 2 co-founders with computer science and engineering degrees and have annual revenues in the range of $5–10 Million are prime for exits. Out of nearly 200 AI B2B startups, only 13 matched the latter criteria. These include:
VC Professionals: Keep this in Mind When Looking for High-Potential AI B2B Startups
Based on the data analysis, as with AI B2B Startups M&A targets, I’d also suggest for VCs to consider the features discussed above when assessing an AI startup investment — (1) 1 to 2 co-founders who hold (2) engineering and computer science degrees. It may be of interest to invest in one of the startups listed in the predicted exit list.
Of course, as a VC you want to get in as early as possible to reap higher rewards from the exit of an AI B2B startup investment. Consider the funding round and total amount of money the startup has raised. It may be interesting to look at their cap table, their investor and board list.
With a bit of curiosity, I re-assessed the current AI B2B startup list for potential VC investments. I found 5 startups that fit the criteria mentioned above but that have annual revenues of between $1–5M. These startups may be an even better investment for early-stage VCs (and for potential economic and control leverage).
Conclusion
I come back to my original question, “what would be most valuable as a VC investing in AI?” I analyzed data (from Mattermark, Crunchbase, and LinkedIn) and identified features that make for a “winning” AI B2B Enterprise startup exit and investment. I identified “winning” features to include:
- Annual revenue $5–10 Million range
- Median of total funds raised $6.5
- Co-founders with engineering and computer science degrees and, at times, PhDs
- 1 to 2 founders
- Median of 5 years old
Using data was a great way of drawing patterns and taking a stab at predicting the future, however, you should take caution with the results. We don’t know if Mattermark’s, Crunchbase’s, and LinkedIn’s data is accurate or up-to-date and it’s hard to say that any quantitative deductions are 100% accurate. That’s why it’s important to use quantitative analysis as only part of an analysis procedure in investments (and in other areas in life and work). Data analysis and/or algorithms today need to be used as augmented assistance and not accepted as the complete truth.
My recommendation is to use a combination of qualitative and quantitative analysis when making decisions. Because after all, it’s not numbers nor quantitative analysis that make for successful companies. It is great people who build great companies.
If you’re a VC fund and either don’t want to do the dirty work or want help doing it, let me do it for you.
If you liked this post, clap 👏👏👏👏👏 and share it!
About Author: Lolita Taub is a TEDx speaker and keynote, a World Economic Forum Global Shaper, an artificial intelligence enthusiast, and an enterprise tech professional and investor at Portfolia. She holds 9 years of enterprise B2B software-hardware-and-services sales experience at IBM, Cisco Systems, and in Silicon Valley. Lolita has been recognized for her work on Forbes, Inc.com, The Huffington Post, Entrepreneur.com, and Los Angeles Times, among other publications.
Follow Lolita on Twitter @lolitataub, visit her here, and connect with her on LinkedIn here.
