We're growing 400% per year, with zero cash left, still optimistic: All the numbers, stats, and details on how we're running our AI startup
I just read an article from Gabriel Mays with a list of "business questions worth asking".
I found the list thought-provoking, however "provoking thoughts" seemed to waste a huge opportunity to actually dive deep by trying to answer, not just ask, those questions.
So there you go. Here's how we'd answer those questions at OneTake:
This month, we want to launch an actual ungated trial so that people can try out the service themselves before subscribing with credit card.
We offer a huge surprise incentive for users to upload their first video (if you do that right after signing up, you get upgraded to the next tier for free) and this results in 77% of new users taking action (and getting to the Wow moment) on the spot.
We run 2-week sprints and release a new version of the software every 14 days.
Our CEO runs a customer demo live for 2 hours to showcase all the new features, answer questions, take feature requests.
20% of all our customers attend every session.
From running close to a hundred live sessions and webinars, here's what we know:
We had assumed our churn would start out at maybe 20%+ month-on-month, like a lot of startups, and would progressively go under 10%.
Instead it's been much better:
Yes, that's our promise.
Many of our customers have never published a video online. Many more have only ever published very little online content because of the "blank page syndrome" coupled with the technical hassle of shooting and editing those videos. We solve all those problems.
So our customers get to finally be content creators and share their message to sell their products and services.
We are in contact with these users every day. We do 1-on-1 Zoom calls with users who have questions, doubts, need hand-holding, etc.
And during our bi-weekly live session, users get to tell us about all their requests and wants.
Our founder and CEO was part of the target market himself (has created thousands of videos and run several education and information marketing businesses) so we started with a vision that wasn't completely possible at the time (the AI tech wasn't yet there when we started in 2021).
We are still pursuing this vision and this roadmap, but we're letting our actual paying users inform the priorities on this timeline and we discard the initially-planned features that seem out-of-scope for what customers want.
A small part of the Product work is staying on top of some of the developments that our customers are hearing from the rest of the market.
For instance, making the assistant "multi-modal" (making it possible to talk to it to give instructions as you would to a freelancer) was a move driven by the competitive space. It was on the roadmap for more than a year, but we pushed it earlier because we want to stay ahead ; remaining "innovative" and moving fast is a big part of our marketing both to current paying users, and prospective customers.
We want to give our customers better confidence in their speaking, presentation, and selling skills.
Allowing them to be "imperfect" at the time of recording is a big part of that.
We also provide free training sessions with experts on related topics : public speaking, voice projection, marketing an advertising.
One way we're doing what's right for the customer is that we grandfather paying users to all new updates, as well to the bonuses offered on later promotions.
This has been a big part of our retention strategy : users know that if they show up on one of our webinars or online challenges and we announce a fast-action bonus for new buyers, all previous users can also request and get access to it.*
* legal disclaimer: This statement is not a contractual clause and should be taken as reflecting the historical record rather than making promises about the future.
That's how we get to grow 10-15% month-on-month, while also increasing retention of the existing users over time:
When you only have a single product to sell, and it's subscription-based, focusing on building something people will love is the only trick we had.
Many of our initial assumptions were wrong. For instance, we initially assumed that we could sell the software the same way that Russell Brunson sells Clickfunnels (as a bundle of a relevant infoproduct coupled with a "free" 1-year subscription to the software), and this offer bombed. Hard.
Selling the software itself was what made the difference:
We have an execution problem. We are taking off like a jet plane (yay, say outsiders looking on) but we should really be taking off like a rocketship.
So what's been lagging?
We think we lack on the product management side. Our CEO does both this and the marketing and the live sessions, etc. So we're shipping reasonably fast, but we could be shipping way faster.
As an example, we were amongst the first in the world to deliver video translation (with lip sync) but another startup put out this feature too at just about the same time. Being one in 2 to deliver a feature in this fast-moving field of AI is cool. But we had a working proof of concept of it 14 months earlier (August 2022)... Despite all the other stuff we shipped in the meanwhile, we think it simply took too long for us to go from proof of concept to product.
We have a very ambitious roadmap, and life is short, so we're right now recruiting a kick-ass Product Manager / Product Owner. If that's you, just get in touch via email!
If you know just the right person for the job, you can also forward them the link below :
Making things accessible to the entrepreneurs who have been stuck because of the technical hassles.
We aim to make something extremely powerful, that fits in a single button click.
We aim to be differentiated. We don't fight competitors on cost.
Our subscription is a higher "step" to join, but this allows us to be so much more generous that we don't nickel-and-dime users on "minutes of video" or "megabytes uploaded" like legacy video software does.
We're going after the antelope :-)
How we do this:
Focus on monthly growth goals (+20% is the goal) and only take marketing actions that can significantly affect the goal.
It might be true in the sense that we've used our CEO's name, reputation, and email list to get started, and this was playing the game in "easy mode".
Getting out to the wider "cold" market is where we're going to see the true strength of what we've built and how well we can sell it.
Right now our CAC is $1,300 (which means the payback period on that advertising cost is 1 year!) so that's way too expensive for our means.
Cutting this down (to under $300) is our new priority.
Opinionated hot take : while there is a LOT of demand from businesses to generate synthetic content (eg avatars such as what Synthesia or Heygen provide), we believe that entrepreneurs will always want to be in the limelight.
We don't believe that synthetic content will replace organic content. It will flood all channels, yes. It will represent an overwhelming portion of consumed content, yes.
But there will forever be an infinite stream of people who want to start a business, share what they know, and be the one delivering that content (whether on stage, in a retreat, in an onlline course, etc.). This is our audience and we're 100% invested in them. Hence our strategic choices.
Well this one hits close to home right now:
So far I (Sébastien, co-founder) have funded the whole venture out of my own pocket.
Now my fingers are starting to feel the linen threads at the bottom of that pocket. And though I'm raising a pre-seed round, it's... the slowest and most frustrating process ever.
So where do we go from now? Well, what we do right now is work on that growth. Go from 15% per month to 20% and then 50% and then higher.
In so many businesses, growth will kill you faster. But here we have awesome unit economics, and it's only getting better:
So every customer we get is a profitable customer:
A year from now, if we've failed, it will be because we didn't reach breakeven fast enough.
That will be if:
So first we'd need to define which role i'd be replaced in:
If I (Sébastien) was fired as a CEO, my replacement would probably reach out to VC networks and raise a proper venture-capital seed round rather than trying to raise from family and friends.
While I'm super happy that I've already raised 20% of my round from family and friends already in the last month, it's a very frustrating process. I like marketing (making an irresistible offer to the market at large to attract the relevant leads so they can self-select and buy) ; I hate sales (having to patiently listen to the excuses and dilly-dallying of one single person at a time, and then followup with neverending patience because humans love procrastinating on a decision).
If I was also fired as the head of Product, my replacement would be much more effective than I am there, just by virtue of being full-time 😅
So we'd finally get a roadmap that is clearly organized in GitLab in a chronological order that makes sense to both the technical and non-technical teams. This would help us move forward faster.
If I was also fired as the spokesperson for the company, my replacement would create more publicly-accessible content. Right now I'm doing a lot of video content, but that is mostly bottom-of-the-funnel : live webinar presentations, live challenge sessions (just did over 20 hours of work of live presentations this week). All of that content is what delivers sales and growth right now, but none of it is "public".
It's not ideal that as a video company, we're not putting out YouTube videos ; it's not ideal that as an AI company we're not publishing videos about AI and the content economy.
If I was also fired as the head of marketing, my replacement would implement more of the "normal" software company playbook. Things that are missing now that should be priority for product marketing:
What's cool about our culture:
* legal disclaimer: This statement is not a contractual clause and should be taken as reflecting the historical record rather than making promises about the future.
What's less cool about it:
We try to :-)
We have a system for the Customer Success team to alert the other teams (especially the dev team) on both critical problems, and customer problems or requests.
A "critical problem" is a problem on the critical path for customers. As a User, you should 100% of the time be able to:
If anything in that chain is broken, we have an emergency (and we have a process for that).
Outside of that, any other bug or feature request is logged into GitLab and shared during team meetings or through our Signal thread.
We probably could do that even more with better product management.
We have a North Star metric for the product (number of hours of video edited per customer per week).
We've acted on that metric (e.g. by sending update notices and reminders to users if we notice that usage is slowing down) but it's not really driving all of our business like it could.
😒 Yep... definitely.
We've done a big shift on this one this year (after receiving advice from Dylan Frost, CEO of The Wholesale Formula) because our marketing used to focus on "online course creators" (our initial narrow niche). It turns out that there are many more businesses that need video for things other than creating online courses! 😉
So we've evolved our marketing to be less restrictive. The total addressable market (TAM) is 5 million businesses in first-world countries that are producing or aiming to produce video content on a regular basis to educate or sell. At our ARPU, that is a $6 billion dollar TAM.
We chose an industry and a business model that has built-in high-retention. Instead of simply editing videos, we also act as the video hosting service and provide features that no other video hosting company can currently match. (e.g. the automatic video translation and automatic playback of the video in the viewer's language). This means we have a high stick rate and pain-of-disconnect, so our retention and LTV will only grow over time.
On this current trajectory:
Right now, it feels like A.I. is everywhere (just like we felt as if the Internet was everywhere in 1999), but that's because it's being "talked about" everywhere while adoption is still in its infancy.
We believe that there are still many more "10x moments" to be had as an A.I. company.
Right now we are surfing the "A.I. trend" because our customers are much more interested about A.I. than they were when I was talking about our artificial intelligence software in May 2022 (honestly ; nobody cared. It was painful ngl 😂)
But we are not targeting an easy AI case that ChatGPT will just take over next month... we're targeting a valuable business case that has existed before AI, and will exist after AI is no longer "fashionable" and is no longer a big search term.
Trends we're latching on:
The world is plunging now into the "Content Apocalypse of 2024".
Because of this, our people (the entrepreneurs generating content) will need:
Nobody has a clue what a world post-A.I.-trend looks like. (Or, even, if we'll all still be here.)
But what we do would still be valuable without the A.I. component. Billions of dollars of video editing software has been sold, and billions of dollars of video editing services have been sold.
This need will go on. The people running [insert big conference or live event] will still want to put the recordings out there. The speaker will still want their keynote edited. Etc.
We are Kittinger jumping down the stratosphere into the gravity well. Bring it on.
To answer this more seriously:
In early 2022 I benefited from a consulting session from Annie Hyman Pratt, CEO of Leading Edge Teams, ex-CEO of The Coffee Bean and Tea Leaf, and author of The People Part: Seven Agreements Entrepreneurs and Leaders Make to Build Teams, Accelerate Growth, and Banish Burnout for Good.
During the consulting day, Annie said something that profoundly shocked me:
"We as entrepreneurs think that our own performance is responsible for 90% or more of the performance of the company. That is false. The real impact is maybe 5%."
I was dumbfounded, so she clarified:
Success comes from having momentum on your side. Too many entrepreneurs are pushing a boulder uphill. If that's what you're doing, you're always going to feel like you need bigger muscles, better discipline, a better boulder-rolling technique.
What you could do instead is find a lighter boulder.
What would be even better than that is to decide you want to push the boulder down the hill.
So Annie helped me see that my business needed a fundamental change. Only 5% of the business performance is my performance as CEO ;so what are the other 95%?
According to Annie, it goes in this order:
And that's how I decided to shut down my 7-figure "training and coaching" business and pick something that I was (1) excited about and born to do, (2) in high demand because entrepreneurs want to create video more than ever, and (3) I could see was just the beginning of a gigantic trend.
(Although, as mentioned above, I was a bit early and had to wait a while for the market to be interested in "AI")
Our vision is for OneTake AI tobe as close as possible to getting a super-talented (human) video editor to work for you.
10x better would mean having "your guy" be both smarter and show more initiative.
If you could somehow have less work to do around the video, that would be 10x helpful. If as a user I could just fire off my phone's camera or my laptop's webcam, record with no fuss, and then maybe just tell the AI what to do with the video once it's ready... Something where I have "zero work left" after I click "stop" on the recording session...
We need to think about this more, but it sure would be an interesting direction to take things!
We're betting that there will be more people wanting to create online content and videos, and more entrepreneurs in particular.
We're betting that if it becomes easier, each of them will create more content.
It's amazing that we've gotten this far, reached $425k ARR, and we're still not really "publicly" launched on all these places that all these other AI startups are visible on.
If it's easy, then once we launch on all these platforms, people will just storm the gates to sign up for our free trial.
I'm expecting things to be hard though, and to require a lot of learning and a lot of focus and har dwork. It's all been hard to get here honestly, so I'm not expecting it to be any different to get to the next level.
Wow, I honestly had a hard time figuring out what this question meant, so I had to go on an epic chat session with our AI bot to answer that... and find out that it's an example of the answer.
Our commodity services:
As a result, we are stockpiling data: the videos themselves.
We plan to do this by making an API available within the next quarter.
Our goal is that any app (for instance a Learning Management System - LMS) that needs user-generated videos can use our software to help their own users get over the hurdle of video editing.
It will also enable users to embed us in all kinds of automated processes through Zapier, Make.com or other connections. If we can allow people to hook us into their Zoom coaching session so that the replay is edited, translated, uploaded to their Content Management System (CMS) or Wordpress blog within minutes, then it's a major win on both sides.
OK, I feel like some of those questions come from Gabriel May's notes after reading some blog post or other, and I have no clue what this one means. Skipping it for now.
Our top 3 biggest opportunities right now:
We're not investing nearly enough on those.
Yeah... so we used an UX firm to make our initial prototype two years ago:
After all the changes and updates we've had, I think it's time we rethink both our UI and our UX.
We have something 100 times more powerful, but we didn't stray too far from the initial plan:
We can definitely improve the onboarding for a new user.
One more of those cryptic questions. I'm going to assume it means "are you making sure that the features you released recently are in alignment with your marketing efforts and your promise?"
And if that's the question, then yes, because we prioritize features like this (in this specific order):
That is literally everything we do.
This is one of the toughest parts of our work on the product, and we haven't perfectly nailed this one yet.
Our entire promise is to remove friction ; that's what people expect from a one-click editor.
But then it's difficult to prioritize between:
So we kind of go back-and-forth. I wish we had a proper rocks-pebbles-sand system as WPEngine's Jason Cohen describes in his great article. We'll get there.
Right now everyone has a simple incentive: to help the company succeed. By succeed I mean survive ; and then, win.
We don't currently have an incentive program.
We have those, and we have a weekly meeting to measure, report and act on these metrics.
Hmm. It feels like we could communicate strategy better.
This is the CEO's role and we need to reconsider how strategy is commincated to the team.
A strategy exists, it's just not as well communicated as could be:-)
Right now, it feels like this for the translation feature. It was costly to build and it is costly to run.
However this gives us the opportunity to have a higher tier, because our generous policy (hinted at above) made it very hard to have a feature-differentiator for multiple plans.
So we're hoping that we can now increase our ARPU with it.
We are probably more cash-efficient than competitors for the simple reason that we're short on cash and almost all significant competitors are VC-backed.
As a result we are probably less effective than competitors in the sense that we have limits and constraints they don't have, so that's a challenge.
Very interesting question!
One thing that comes to mind is that we're paying a lot of money for GPUs right now, and when they're not running our own service, we might be able to "re-sell" this GPU time on vast.ai or similar. Maybe.
We celebrate by having a live session called "The Real Time" every other Friday with customers!
Recently we ran mostly to our French customers because US-based customers were much fewer in number and didn't attend the English-language live sessions as much. We want to go back to having an English session again. This year we've also run a special session in Japanese too!
We haven't done an evergreen webinar because the software evolves too fast: by the time the evergreen webinar is up, some of the objection-answering slides from the presentation are obsolete because we've solved that objection with a simpler, better feature.
Right now all our webinars are live and that's where most sales come from.
By building a webinar that is more evergreen, now that the product is solid, we can definitely focus on 20% of the work (attract new signups and trials) and potentially still get 80% of the conversion we get on the live ones.
Wow, answering all those questions was a really interesting process. It brought us some new insights, as well as a very useful deep-dive to share with the team so everyone can better understand where we're at and where we're going.
I hope you will also answer some of those questions, and share it with us!