Yesterday I sold my 8% position in Datadog (DDOG) and put the money into a company I have thought about but never held before, Braze (BRZE). Here’s why:
Datadog was already my smallest position. Then all the debate about it here made me realize that I don’t understand enough about the product to judge whether it is headed for greatness or irrelevance. Despite holding it since Q2 of 2021, I have never understood its “secret sauce,” so with yesterday’s strength, I exited.
The debate about DDOG and the parallel discussions about AI/ML happening everywhere made me wonder if there might be another company already at work with that technology whose advantages I could better gauge and that didn’t come saddled with the geopolitical risks that made me so sad to exit NVDA.
Braze has come up on this board a couple of times since their IPO in Nov. 2021. Replies to this thread by nlieb link to other mentions of the company on the board. I looked at Jamin Ball’s last Clouded Judgment post and Braze didn’t seem to be outrageously priced at the moment, so I decided to look more closely.
What do they do? From their IR website:
Braze is a leading comprehensive customer engagement platform that powers customer-centric interactions between consumers and brands. Our platform empowers brands to listen to their customers better, understand them more deeply and act on that understanding in a way that is human and personal. Using our platform, brands ingest and process customer data in real time, orchestrate and optimize contextually relevant, cross-channel marketing campaigns and continuously evolve their customer engagement strategies. 1,770 customers around the world trust Braze with their most valuable assets: their customer relationships.
In other words, they facilitate brands communicating directly with their customers. Anything that helps B2C interaction is a well-established trend right now. As the CEO, Bill Magnuson put it:
I think that the secular trend toward being able to build up first-party relationships inform those first-party relationships with first-party data and then take action on them directly is one that’s becoming an imperative for brands of all kinds across every vertical and of every size.
A personal example of how valuable that is. In addition to being clergy, I am the author of eight books. Four were published by commercial publishers and four by my prior employer, where they were done as works for hire (i.e. as a non-profit, they get all the royalties; I got my salary for writing them). In both kinds of publishing, the majority of people bought the books on Amazon.
In neither case could I get any information on who bought my books. The closest I could come is the map Amazon provides to authors to show how many books were sold in a given region. But if I wanted to get people to subscribe to a blog or get people who bought a previous book to buy a later one, I had no access to that data or even to them directly with the IDs removed. That was true even though, for the works for hire, I also served as the publisher (as the ED of the non-profit for which I wrote) and we published the books ourselves through Amazon’s publishing service.
As the publisher, I could get the number of titles sold and the $$ that gave us. As the author, I could log into my author account, look at the map and see what regions they came from. But the thing of real value? The information on who the actual customers were or some opportunity for direct contact–customers who might have become donors to the non-profit or who might want another book in the series? Nope. Nothing.
And, as the publisher, I didn’t have access to the author map except by virtue of being the author myself. That data was vital in developing our marketing efforts and being able to see, for example, whether a mass mailing to one region caused a spike on the map. But had I not been both publisher and author, even that data would have been walled off, each from the other.
So I understand why what Braze does is valuable, and why facilitating B2C access is a secular tailwind.
Why Braze and not others? Braze is certainly not alone in this space, with Adobe, Salesforce, and Oracle being the best-known competitors. Here’s the Gartner Peer Insights for the category.
To get a better handle on this, I listened to the latest earnings call. Braze does not report on Q1 until June, so the latest call is from Q4 of their FY2023 on March 30. The recorded call is here. The only transcript I could find that wasn’t behind a paywall is here, (where Braze is routinely transcribed as “Brave,” “base,” “Brac” or even “bra.” You get what you pay for, I guess.)
Like with some of our other companies, there was talk of customers replacing several “legacy” products with the single Braze platform. Of most interest to me, however, is that the company was founded by two engineers. Bill Magnuson, the CEO (97% Glassdoor rating), and John Hyman, the CTO. From the transcript (note that in all of the transcript quotes below, I have corrected the company name or other transcription errors that I found):
“In 2017, we invested in a dedicated team of data engineers and data scientists focused on using machine learning to build AI into the product.”
So they are not a company currently scrambling to get in on the AI craze. It was built in from the beginning. And they have kept up…
In May of 2022, we started investing in generative AI as a marketer copilot, building GPT-3 into Braze for AI copywriting which saves customers’ time when creating subject lines and messaging for their campaigns. Last December, we integrated DALL-E into our image library and customers have been using both very effectively to be more agile, speed creative production and more easily test and optimize content for different variants. We’re very excited about the potential for these tools to enable even small marketing teams to wield an immersive creative vision. While also making it easier for them to take advantage of Braze’s sophisticated and differentiated testing capabilities by lowering their creative production burden. We have also already seen examples of ChatGPT successfully leveraged to generate valid message templating syntax and to speed along integration with source code samples.
Later in the Q&A, Magnuson spent a lot of time talking about AI and the difference between what he calls the “co-pilot” approach above vs something fully autonomous. What impressed me there was his nuanced approach to AI/ML and the careful consideration of ethical implications and how damaging errors in communicating with customers without any human interface can be.
While Zuckerberg funneled human lives into his “move fast and break things” mantra, Magnuson is careful to distinguish the interactions that can be fully automated without worry, contact that needs a human “co-pilot,” and interactions that lie somewhere in the middle.
Toward the end of the call an analyst from JPM Securities asked Magnuson about the open letter from the Future of Life Institute asking for a pause in giant AI experiments. Did Magnuson sign it? Would he? [My note here: The letter came out on March 22 and this is a call on March 30.]
Here is Magnuson’s full response:
I was preparing for earnings and so I didn’t dig into that yesterday. But I definitely think that we — with all technology that could go either way in terms of being helpful or becoming dangerous that we should be responsible around the development of it. I think that Braze is a place that’s always been really measured about what AI is capable of and where we should trust it. I talked about that earlier in the sense that when you’re interacting with first-party data and first-party relationships, the bar is just a lot higher for when we should go and trust the AI. And you’ve seen that Braze has never been one to promise a Magic Black Box. It’s also all your problems. But it’s definitely an exciting time for AI development.
I think we need to continue to be mature and responsible about how we do it. I think that there should be more openness in terms of how these large models are being trained, what kinds of rule sets are being applied to them, what kinds of data sets are going into them. And I think that without additional transparency, it’s going to be hard to trust continued development of it. And so I definitely support continuing to put pressure on this ecosystem. It’s not something that we should allow to just like run unchecked.
I really, really liked that answer and find it all too rare. While all risks to AI development can’t be known, a CEO with this kind of attitude moves the risks associated with AI and ML to a much lower and more comfortable level for me.
What about the numbers? Braze is a small company that is not yet profitable or free cashflow positive, although they charted a course in the call to get there in 2024 (their FY2025). They did note that prioritizing profitability will likely mean little to no acceleration in revenue growth from this year to next. Not little to no growth–little to no acceleration.
Bearing in mind that this was the report of their first full year as a public company, here are the top-line paragraphs from the CFO, Isabelle Winkles:
We reported a strong fourth quarter with revenue up 40% year-over-year to $98.7 million. This was driven by a combination of existing customer contract expansion, renewals and new business. Our subscription revenue remains the primary component of our total top line, contributing 96% of fourth quarter revenue. The remaining 4% represents a combination of onetime configuration and onboarding fees as well as other professional services. Total customer count increased 29% year-over-year to 1,770 customers as of January 31, up 395 from the same period last year and up 55% from the prior quarter. Our total number of large customers which we define as those spending at least $500,000 annually grew 46% year-over-year to 156 and as of January 31 contributed 57% to our total ARR compared to a 52% contribution as of the same time last year.
Compared to last quarter, this reflects a 5% increase from 148 large customers that contributed 56% of our total ARR as of October 31. Measured across all customers, dollar-based net retention was 124%, while dollar-based net retention for our large customers, those spending at least $500,000 annually was 126%. And Expansion was broadly distributed across industries and geographic regions. Revenue outside the U.S. contributed 43% of our total revenue in the fourth quarter in line with the prior quarter and up from 40% at the end of fiscal year 2022. In the fourth quarter, our total remaining performance obligation was $456 million, up 22% year-over-year and up 11% sequentially. Current RPO was $313 million, up 31% year-over-year and up 10% sequentially.
The year-over-year increase was driven by contract renewals and upsells and the signing of new customer contracts. Overall, dollar-weighted contract length remains at approximately 2 years.
43% of their revenue comes from outside the US with the second largest office in London. They just acquired NorthStar, a company they had been working with for several years to expand operations in Australia and New Zealand.
They are in many verticals: retail and e-commerce, media and telecommunications, and now moving into travel and hospitality. All verticals that are, themselves, growing and transforming.
An analyst from Cowen asked the following question of the CFO that made me ask a question I’ve not yet been able to answer.
But just you guys are a bit unique out there. You’ve got a pretty diversified pricing model, other companies. There’s companies with seat-based licensing models that have seen pressure because of headcount growth. There’s other companies with consumption models, seen pressure because of optimization. Again, you guys are much more diversified. Do you think that lends to more durability and net expansion rates. And I guess if not, what parts of the model could come under more pressure given the macro trends.
The answer was around the various pressures that we’ve been hearing about in other calls. But the question implies that Braze has some kind of hybrid pricing model or optionality when it comes to seat-based vs consumption models.
The answer on the call didn’t focus there and I wasn’t able to find out their pricing options/structure. But as we have gone back and forth on this board about the pros and cons of each model, I found that reference to Braze somehow having both to be an interesting distinction. So if anyone can figure out why Cowen thinks they are unique in their pricing, I’d be interested to know.
As I’ve said before, I began investing with TMF in early 2021 and with this board later that summer. So I jumped in when everything was priced to perfection and lost quite a bit in the rout that followed, without any cushion from prior years to ease the pain. Across 2022, it felt more like trading than investing, as people jumped in and out from quarter to quarter. I’m done with that part. Too stressful; too costly.
In taking a new position in Braze, I’m not thinking that I have magically caught the company just as it’s about to ride up the S curve. I have taken the position because I think they have built a company well-suited to take advantage of the tailwinds of increased B2C interest, the transformative power of generative AI, and the move to consolidate operations in single-platform, cloud-native solutions.
Braze’s offering is mission critical–to help brands communicate with their customers directly and at scale. Their management has expertise in the technology, saw the possibilities for AI in their business years ago, and would rather get it right than get it fast (which is also how I’m feeling about my own portfolio right now).
They are asset-light with recurring revenue; and 40% growth in the year of our Lord 2022 seems pretty good to me. They’re willing to slow growth to become profitable, which seems wise given the uncertainties of the macro.
Lastly, unlike Datadog, I understand the business and the value it provides to customers, making me better able to evaluate whatever innovations, acquisitions, or other modifications they make to their business down the road. So, I’m in. Hopefully, they’re Q1 report next month won’t prove me a fool (small f).
My portfolio (as of today’s close) looks like this:
Then there’s just over 1% in AI, which I hold for my brother.