Intercom’s "top to bottom" AI bet 🏊♀️
Making sense of a potentially category-upending trend, segmenting churn based on a product’s vision, and countering the harshness of the fundraising process.
Welcome to the 73rd edition of The SaaS Baton. A fortnightly newsletter that brings you hand-curated pieces of advice drawn from the thoughtful founder-to-founder exchanges and interviews taking place on Relay (curated with 💛 at Chargebee for Startups) and the interwebz. So, stay tuned!
In this edition, you’ll find the following instructive and inspiring pickings:
#1: Intercom’s co-founder and CSO, Des Traynor, estimates the unprecedented effects of deeply integrating generative AI — “a whole new muscle we’ve never had to build” — on their product strategy, cost structures, and customers.
#2: Maze’s founder and CEO, Jonathan Widawski, on a churn/retention perspective shift that made them recognize how certain usage patterns and cultures of adoption should merit an altogether different plane of attention than others.
#3: Chezie’s co-founder and CEO, Toby Egbuna, distills notes from the devotedly insistent course — past 70 rejections — of their fundraising efforts.
🗞 Recently on Relay:
Heuristics and Hunches (May 19th) — Pursuing an Open Source, Bootstrapped, Long-Run Path towards Serving the Fortune 1 Million with Typesense’s Co-Founder, Jason Bosco
— Why open source is an ideal bet for developer tools
— How the SMB market is (predictably) left stranded
— De-risking a flaw that both VC-led and bootstrapped paths share
— Monetization models: open source vs. open core
— Figuring what to build next with the Typesense community
Heuristics and Hunches (May 12th) — Why “Charge More” is Way-Easier-Said-Than-Done Advice, and Other Data-Driven Pricing Lessons with Uizard’s Co-Founder, Tony Beltramelli
— Being data-driven even when pricing is really an art
— Measuring perceived vs realized value (and pricing’s PMF signal)
— Going beyond standard segmentation
— Avoiding competition-based pricing without adequate research
— Going down the usage-based pricing rabbit hole (and deciding against it)
— The (limited) value of working with consultants
— Why changing prices often is hard and what they’ve done instead
AMAs (upcoming!) — I’m A Smart Bear (Jason Cohen), founder of 2 unicorns, both bootstrapped & funded; bought, sold, and invested in startups. AMA!
#1: Intercom’s “top to bottom” AI bet
(From: Intercom’s Des Traynor) (Source: The Generalist)
Honestly, it is a new variable. It’s not something we’re used to in software – there’s basically never a time where you come up with a good idea and say, “We can’t afford to build it.” That’s just not a thing, right?
That changes when you’re building with AI. There are features we could build that we won’t because they’re too expensive.
For example, we could use GPT-4 to summarize every conversation that every customer has with every business on Intercom.
We could do that, but it would cost a lot of money because Intercom powers 500 million conversations a month. That’s a lot of API calls, right?
It requires a different kind of thinking for us. Just because a feature is a brilliant idea doesn’t mean it gets shipped. You have to think of what it might cost at Intercom scale.
…
It is a new concern for our margins though within the context of the broader industry, relying on an AI model isn’t prohibitively expensive.
Depending on where a support agent is and how complex the issue they’re addressing is, it costs between $5 to $25 per conversation, and each agent is tasked with closing between 50 and 100 conversations a day.
The cost of calling an API is a rounding error compared to a company’s fully loaded expenses.
My take on it is: yes, there might be an application for margin here, we don’t know what it is, and the ground is moving very fast.
Second-order concerns will start bubbling up, where the different model providers begin undercutting each other and driving down the cost of usage. All of which is to say that I don’t think there’s a way to be particularly smart about cost right now.
You just have to build things that are deeply valuable and be careful about shipping features that are cool but could incur massive costs at scale.
…
Intercom is already a lot of the way there. The challenge is going to be articulating AI’s value and shortcomings to our customers. They need to learn how to feed it better, how to give it better context.
It’s about helping customers get the best value out of these technologies and also understand what’s going on in their business.
Right now, you turn this thing on and say, “Well, I hope everything’s good!” You can watch thousands of conversations fly by, but that’s not an easy way of staying on top of things.
Traditionally, a support leader might walk the floor and hear what’s happening from their team. Maybe there’s a big bug that’s leading to a spike or shipping’s delayed.
We have to do a lot of work on our end to surface that kind of information – to say, “here’s what’s going on” and share that on behalf of the bot. That’s a whole new muscle we’ve never had to build.
Related Relay read: Vue.ai’s founder, Ashwini Asokan, on the challenges of navigating the many unknowns of AI-first B2B products
#2: The churn you (don’t) want to improve
(From: Maze’s Jonathan Widawski) (Source: CHURN.FM)
Churn has always been a key question mark at Maze. How do we compute churn? What’s the right way to think about churn on our end?
So I hate talking about churn as the end-all, be-all [thing]. There’s no solution to churn. Churn is something that is bound to happen within your product.
I think, for us, it’s been about what is the natural usage that people will get out of our product. And how does that change across different conditions, different industries….
As we started getting a better understanding, it was important to know which type of churn we really wanted to improve and which type we didn’t really want to improve.
Internally, what’s interesting is, we have this sentence that we use all the time: Which is, at Maze, we win when testing and research is seen as a cultural shift within the organization.
And we lose when testing and research is seen as an ad hoc activity. What that really translates to is that there are 2 [types of] usage that we see it Maze. There’s usage that’s purely transactional.
[For example] I use Maze as a one-off testing platform when I need to run a test for smaller companies or companies that don’t have the level of maturity to embed research into their processes.
Then there’s the key example when we actually change the culture within the organization, which is when we’re looking to make research something that’s happening at every point in the product development process.
So, we could look at all of this [as] blended. Which is what we used to do for a long time. We were like, ‘how do we solve this problem? How do we solve for this number we don’t know how to impact?’
Then when we started separating these 2 types of churn, we started looking at the real picture. And you want to improve it for people who’re actually successful users of your product.
With the other type, you’re going to have a much, much harder fight. You’re going to have to go against the natural frequency of usage of your product for teams that don’t have that maturity level.
And getting people to that maturity level is more than product usage. It’s more than creating great templates….It’s a lot of work.
For us, simply splitting these 2 and looking at how we can improve for the entities that we wanted to serve was really a game changer.
We had a conversation with Andrey, the CEO of Miro, about this. And they have something very interesting to say about this as well. Which is, at Miro, the way they look at churn is that, for them, success is collaboration.
So they have paying customers that are teams of individuals, people that just want to brainstorm within the product. Even though, this revenue is technically SaaS, but if you aren’t 3 people using the platform, Miro doesn’t consider the revenue as SaaS revenue, or recurring revenue…
For us it’s almost this shift as well.
Which is how do we think about this revenue. Does an individual using Maze for a company below 100 people, is this actual MRR? Or is it going to churn so fast that it doesn’t really matter that we count it as such?
It’s a perception shift.
And [it] will allow you to know where you want to solve for churn.
Finding these hard-won founder takes valuable? Consider forwarding this edition to an ever-curious teammate or a much-cherished SaaS friend? :)) New readers can sign up here.
#3: “Pitch vision or traction but never both”
(From: Chezie’s Toby Egbuna) (Source: LinkedIn)
Fundraising is tough. Very tough. It took us 70 no’s before we got a yes.
Now that I’m through it, I wanted to share some tips I wish I knew before I started:
1. Pitch vision or traction but never both
A few months into our raise, we were stuck. We’d gotten over 40 no’s and my morale was at an all-time low. I got connected to [Vanta’s] Christina Cacioppo, and we exchanged a few emails. Christina hit me with a gem that totally changed the way that I approached raising: pitch traction or vision, but never both.
At the pre-seed stage, since you probably won’t have much traction, sell your vision to investors. Talk about what the world looks like if you’re successful and your game plan to get there. When I started doing this, I could literally see the expressions on investors’ faces change; they became much more interested.
2. You won’t change an investor’s mind
Investors tend to have a clear idea of whether they want to invest in your company within the first five minutes of the pitch. If they are not interested, move on and don’t try to convince them otherwise. There is little chance that you’ll be able to change their mind, and you’ll waste valuable time that you could be using to find other investors.
3. Get the first check asap
One of the first questions that investors will ask is “who else is in?” If you can get some commitment before fundraising more broadly, it will give you an edge. Try to fundraise privately and get some sort of commitment. Even if you’re only able to get 10% of the round committed, when you start formally raising, if you can tell investors that you have only been fundraising for a week and 10% of the round is full, they’ll be more interested.
4. The door is never truly closed
For the investors we spoke to that passed and asked to be kept int he loop, I added every one of them to our monthly updates so they can track our progress. If/when the time comes for our seed, they’ll have 12+ months worth of updates to refer to.
5. When investors show you they aren’t interested, believe them
If you find yourself following up with an investor two, three, or four times, they are probably not interested. We got commitments from our investors within 3 weeks. If you find yourself following up multiple times, they probably aren’t interested, and you’re better off finding investors that are.
6. The best way to get a meeting is still through a warm intro
Although cold emails and filling out Airtable forms can work, the best way to get a meeting with an investor remains a warm introduction. It’s an antiquated, inequitable, and lowkey racist practice, but it’s an unfortunate reality. If you know someone who can introduce you, use it to your advantage.
The entire fundraising process needs to be redesigned, but until then, this is what helped me. Hopefully, this helps another Black or under-represented founder who’s going through the process.
Related Relay read: Pipedrive’s co-founder, Timo Rein, on the many hiccups and difficulties of raising venture capital
🤝 Founder social:
Until next time,
Team Relay (Chargebee for Startups)