What's 🔥 in Enterprise IT/VC #366
Reflecting on 27 years of enterprise software investing + understanding the experience trap
As we prepare for our AGM or Annual General Meeting for our investors, I had a chance to think about what I’ve learned in 27 years of investing in enterprise software startups, and here it is👇; also applicable to founders.
I can remember impostor syndrome at its strongest in 1996 when I was interviewing a super talented and experienced VP of Sales for a portfolio company who was 20 years my senior and thinking what the hell can I offer this person and how was I in this position. So, I wore glasses to look older, I asked a lot of questions that I learned to ask from reading books, but that feeling never really went away. I just couldn’t wait to have more experience so I would never feel like this again.
My mentor and partner, Bob Lessin, though, would always remind me, our fund Dawntreader Ventures, speaks two languages, the corporate language and the Internet one as well.
“That is where you come into play, Ed, everything is so new, you have the opportunity to define yourself and in many ways the experiences of the past can hinder building in this new world.”
I took that to heart, and it’s an idea I continue to absorb. After 27 years of investing, I do have some experience that I so longed for, but I can’t also let it blind me, and stop me from believing in the future and how things can be way different. Trust me, it’s harder than you think. Being cognizant of this is first and foremost. Surrounding yourself with youth in mind and age can also help you continue to challenge yourself which is why I love my team at boldstart.
To that end, I’m super optimistic about the future and super excited to continue to meet new founders who are just a bit crazy and ready to create new markets and new ways of doing things.
In enterprise software, it’s the same shit that continues to recycle in patterns. We will always need better developer tools, software to observe and keep lights on, better security, productivity software, and software for customer facing roles from sales, marketing, and customer support. These needs will never change, people just find another way to 10x what was already built.
As always, 🙏🏼 for reading and please share with your friends and colleagues.
Scaling Startups
Another banger from Lenny’s Newsletter: Lessons from an enterprise startup failure - this is one mantra I will continue banging the drum for - you have to zoom in to the individual user’s day to day life and how your product will make their lives infinitely better, need to go niche to go big, zoom in to zoom out - more here
We defined our ICP as nontechnical business analysts using Excel to crunch big data sets. In retrospect, that was much too broad, particularly because it said nothing about why they were using Excel or the value of the problem they were solving with it. So we ended up with a large basket of ICPs: a scooter company used us to manage location data, an HR team used us to manage employee data, and a retailer used us to analyze distribution. We didn’t really have a framework for accepting one and rejecting another, so we tried to support them all.
If we had a properly defined ICP, we would have seen it start to fray with each new use case. ICP fray confuses every decision for a very simple reason: to know what problem you’re solving, you need to know who you’re solving it for.
Another practical, how-to post in the boldstart ventures Developer First Founder Toolkit from our Operating Partner, Anna Debenham where she was early product leader at Snyk as empolyee #4 to scale
👇🏼 This is powerful and a master class of running a startup, well, during a war
If you missed my post last week on the New Era of Inception Investing and the race to be first, here’s more data; this fits into my framework as the Inception Jumbo Round, and it’s only going to 📈.
SDR Math in the future - lots of truth to this, need to try different playbooks, and tools as well like Clay (a port co)
🔥 must listen on Software Snack Bites by Shomik Ghosh (my boldstart partner). He hosts friend Ethan Schecter who was the first VP Sales hire for Snyk back in the early days of selling to first few customers and still at Snyk as it has scaled to over 2,000 customers - so much practical advice from Mr. “Trade 💰 for Speed” Schecter!
So true…great video for all the VCs out there - thanks Nakul Mandan for sharing
Mike Moritz on what made Sequoia so successful? "I dont think its very complicated. Its grit, persistence, tenacity, determination, refusing to give up, not tolerating low performance, insistence on team, very simple goals and showing up for work every day."
YouTube link here
Enterprise Tech
Another reminder - PLG is a continuum and a great top of funnel lead gen mechanism but eventually as you scale, all roads lead to enterprise and 🥩 dinnahs
Super impressive milestone from Cribl which crossed the $100M ARR mark in < 4 years 🤯 - started out making it super easy to reduce costs for Splunk and has expanded over time - go niche to go big!
We're proud to share an incredible milestone: Cribl has surpassed $100 million in annual recurring revenue! We achieved this feat in less than four years, making Cribl the fourth-fastest infrastructure company to reach $100 million in ARR.
Cribl does differently. From our founding, we’ve looked at the market from a first principles perspective and sought out opportunities to build something entirely new. We find the hardest problems our customers are facing that are not being met by other companies, and we go after those problems aggressively. When we started this journey back in 2018, it was clear that IT and Security teams needed to process data differently. The one-size-fits-all products on the market didn’t meet their unique needs and we viewed this as an opportunity to build fit for purpose products for IT and Security.
That’s why we built Cribl Stream, which has empowered customers to more effectively and securely manage rapidly growing volumes of data. When we first launched Cribl Stream, we were a small company but we knew we had a product that was something very different than what was in the market. Customers knew that what got them through the past decade was not going to get them through the next decade, and Cribl Stream gave them a new way to collect, process, and route their data. We quickly realized how powerful it is to give customers choice and control over their data––something that was missing in the market and provided a ton of value to customers.
Taking a different approach continues to inform how we build products to this day. Our newest product, Cribl Search, flipped the legacy search model on its head. We gave customers the power to search data in place. No longer do customers have to collect and store data before they can analyze it. Instead, they can search for the valuable nuggets without the cost and complexity of having to ingest all raw data. As we look ahead to our upcoming products, we remain committed to doing different to meet our customers’ unique needs.
Reminder, open source AI models are coming for closed source - rate of improvement is dramatic, more efficient and of course more control…
However these super important open source models can just as easily be killed with too much regulation - here’s Andrew commenting on the new Biden AI Executive Order (Axios has a nice overview here)
Reminder for any AI-first startup (whatever that means) that OpenAI will be coming for you…
Growth is dead but there are pockets of early growth, Series Bs, for example in enterprise infrastructure land that are still getting phenomenal rounds done. Here’s a clip from the recent podcast I did with Harry on this topic.
Case in point 👇🏼 - Chainguard in the software supply chain security market raised a healthy $61M Series B just announced this past week.
Chainguard–the leader in software supply chain security–announced it has completed a $61 million Series B round of funding led by Spark Capital, bringing the company’s total fundraising to $116 Million alongside accelerated growth for its Chainguard Images solution
💯 Hybrid for the win IMO - speed on desktop but cloud great for collaborating with teams, version control, governance…from Kelsey Hightower
If laptops keep getting faster, cloud based developer tooling is going to become less appealing. These machines are so powerful that most of the value add can run locally. Maybe we'll get a hybrid model, SaaS running locally, but it's clear where things are heading.
Markets
Little margin for error in markets today - Confluent down 44% in one day after earnings this past week - “Confluent’s stock gets hammered on weak revenue guidance” (Silicon Angle)
Dig Security in DSPM space rumored to be bought for $300-400M by Palo Alto Networks after having raised $45M
Dig Security has developed a platform to prevent data leaks for databanks in the cloud. The company was founded in 2021 and has raised $45 million since then. In its most recent financing round in 2021, Dig Security had a company valuation of $125 million and it is now being sold well above that valuation at between $300 million and $400 million. Investors include Team8, for which this would be one of its most significant exits in the current wave of acquisitions of cybersecurity companies. Other investors include CrowdStrike, CyberArk and Merlin Ventures. Later investors who will enjoy lower returns include Signal Fire and Felicis Ventures.
🤔 valuations today from Gokul Rajaram and Jason Lemkin
Freshworks is a SaaS star. $600m ARR, profitable, 20% YoY growth (adding $120m ARR annually!).
Enterprise value $4b (market cap of $5.2b minus cash and short term current assets of $1.2b).
The market is assigning it less than 6x multiple on forward revenue.
Think of what this implies for a private company to be valued at $1b.
I think we are going to see a shift to running more workloads locally on laptops in the workplace. Apple has been way out in front here, with the M2 processors, which are well equipped to run machine learning models. There are big advantages to running the AI locally. First, its more secure. Admins can push out models to users based on LDAP group using the endpoint management software. Different employees can get different AI models based on need and authorization. Second, this will be cheaper. Since inference is most of the cost of running AI, why not take advantage of hardware you're already purchased, rather than the pay as you go model with GPU's in the cloud? Third, the experience is better because you don't have the latency of pinging every request up to the cloud and back. I work in AI, and for development we train and run models from repositories like HuggingFace locally on our Macbooks al the time. For training our largest models, though, the cloud still makes sense. These simply require too much data to store locally. You need some kind of big data platform, even if its just S3, to process and store the huge quantities of data. It will be really interesting to see how IDE's involve given these constraints - you want to do as much locally as possible, but all the big data processing tasks still belong in the cloud.