49
34 Comments

I created a product with 2 friends and sold it to Microsoft for 8 figures. AMA!

Hey all,

I created Lobe (https://lobe.ai) with two other friends. The goal was to make it as easy as possible for others to learn about ML and create their own projects with it. I learned a ton, met a bunch of really incredible people, and had a great experience throughout the ~2 years to took to build, ship, and sell it.

In the end, we sold to Microsoft for a life changing, now I can retire if I want (I'm not, onto the next project) / easily live off interest type of deal. Microsoft was a great home for it and they've been trying their hardest to be good custodians of the project.

How can I help you?

  1. 10

    I have no question to ask, but I just want to say that your landing page is one of my favourite of the last few years. I said it when I first saw it and say it whenever someone brings up Lobe in convo. It's just so crisp, clear and simple!

    Actually here's a question then - who designed it? 😅

  2. 5

    What did you learn about negotiation during the acquisition process? Anything you'd do differently if you could go back, or tips you'd give others negotiating a similar acquisition?

    1. 2

      Two tips.

      Competition, and People.

      For the acquisition price, as much as possible make sure to have others in the loop on the process too. Competition is your friend.

      For the overall happiness in the deal (sometimes a completely different question to the one above), make sure to really, really, get to know those involved. Who will you be working with day to day? What are their goals? The people you work with can make all the difference for the 2-4 years you stay at the company.

      To a large degree I failed miserably in the people department. And did pretty okay in the competition department.

  3. 3

    Hi Adam,

    I had a look at lobe.ai and I must say I am super impressed.. It's insane man.. Super cool product.. Will definitely try to use on some personal project. A quick question though, I just did a quick google search about Lobe.ai and found this article https://techcrunch.com/2018/09/13/microsoft-acquires-lobe-a-drag-and-drop-ai-tool/ and this is a 2018 article. So did the acquisition happened in 2018 or now? Actually it doesn't matter, the product is super awesome, but just curious!!

  4. 2

    Wow, this is soooo dope!
    I've definitely bookmarked it, at for sure at some point I'll use it one way or another!
    And yeah - I love your presentation video! I mean, how easy you make it to train a model is insane.

    The only question I have: I see "object detection" is coming soon. Any rough ETA? (since of course, I'm quite interested in that 😁)

    1. 2

      I see "object detection" is coming soon. Any rough ETA? (since of course, I'm quite interested in that 😁)

      Haha oh man you and me both! I did my time and have left to work on my next project, but I know the team's working on it and I've got at least two projects myself that I'll use it on.

      1. 1

        I thought object detection is already solved. Am I missing something?

      2. 1

        Hehe, good one! Lets hope it'll be soon 😁

  5. 2

    How do you deal with the entrepreneur roller coaster?

    I imagine you had some lows before the sale, how did you push through?

    1. 2

      How do you deal with the entrepreneur roller coaster?

      Having really solid co-founders was key. It got hard, there was a ton hours/sweat/etc put in, and having others who were going through the same thing and who you could talk to was great.

      I imagine you had some lows before the sale, how did you push through?

      Actually we didn’t, and again not sure this is exactly how you should build a company, but we stayed heads down and when we launched had more interest than we could possibly handle. Part of why we went to a bigger company, way more resources.

  6. 2

    This is inspiring, congrats! Two questions, if you don't mind:

    • What was the most difficult part of building the product?
    • Do you have a ballpark of how many hours went into building it?
    1. 2

      What was the most difficult part of building the product?
      Oh man, let’s see…

      It might have been the amount of iteration we did on the product. We’d build something, decide it wasn’t quite right, and then mix in up again. It took a ton of time.

      It might also have been the level of polish we wanted in the product. Honestly there’s a MVP version of the product that could be a basic script, a few hours worth of work. All the iteration, animations, etc., the stuff that changes a product from being okay to great, made everything take longer, but it paid off in the end.

      Do you have a ballpark of how many hours went into building it?
      Hum, well we roughly all spent every day working, no weekends or holidays, for a little less then two years. So:

      1.7 * (365-20) * 8 * 3 == ~15000 hours.

      Probably a bit less, but idk, roughly that.

    2. 1

      This comment was deleted 3 months ago.

  7. 2

    wow Congrats! I'm a little confused about your product though. It doesn't seem to be monetized (currently). Seems you just get on a beta list? Can you share some numbers to back up the exit amount?

    1. 3

      Yeah great question.

      So I’m not totally sure I’d recommend every company be built this way, it’s just one of the options. I also bet it's pretty different from most of the other stories here. So take this with a gain of salt:

      We decided early on that the goal was an interface one. The problem was tensorflow is way too high a barrier of entry for most folks, so what could we do to make it easier for “regular” people. Architects, designers, etc. Folks that are generally smart, but don’t know the entire tech stack required to get started.

      With that in mind, we focused on getting the interface right first. That was the problem to solve. We even turned down paying customers in the beginning. The thought being that in the end if we had a way that made this super easy, we could monetize it. We also wanted to keep it as free as possible, if our goal was to make this easy for regular folks, putting a big price tag in front of the product flew in the face of that. Charging big companies to use it but letting little guys use it for free is pretty ideal, at least for this type of idea and goals.

      And to that end, it is monetized. For Microsoft, that means a "deploy to Azure” button in the app. There maybe other ways incoming, too.

  8. 2

    When did you start Lobe? What were the breakdown in responsibilities between you 3 as you managed the company? Did you build Lobe with a 9-5 or was this your full time focus?

  9. 2

    This is insane man 🚀 congratulations!! I'm currently building listnr.tech (to make podcast easier for the masses)

    Just two questions -

    1. What was the entire acq. Process like? Did MSFT reach out to you/how did they find you etc?
    2. how much did you sell for? 💵👀 (Feel free to ignore this one hahah)

    Best,
    AB

    1. 3

      Thanks! Good luck on listnr.tech, I've always found it crazy how much work needs to go into setting up a podcast. I've had some friends who've found it enough of a barrier to just not do it. Another problem I've heard others complain about is the lack of connection with their audiences. If you're broadcasting out on other platforms (Facebook groups, YouTube, twitter, whatever) you get a ton of data back about who your audience is and when and how they participate. But with a podcast (for better or worse) seems like you really only easily get back download numbers, maybe some rough locations based on IP, that kind of thing. Nothing about who they are, how they listen, how much they listen to, or why they're there. Not sure what the right, possibly privacy centric way is to solve this, but anyway it just came to mind.

      1

      It was a wild ride, several companies reached out right when we launched and ultimately Microsoft was the right call. That was the best part, aside from one or two intros from our VCs, the best offers came from folks who reached out on their own. In Microsoft's case, Jaron (https://en.wikipedia.org/wiki/Jaron_Lanier) was the first to make contact. As a little side note, he's a cool person, if you want to go listen to a few of his talks.

      In general I think the best thing we did was stay quiet for a long time and just keep our heads down building. By the time we launched we had something we were proud of and could move forward from there.

      2

      Haha oh it was a great offer :)

  10. 1

    I am downloading Lobe as I write this ! Super awesome Product.

  11. 1

    how do you position a startup to sell for that $$? did you have any prior connections to microsoft? or years on industry relevance?

  12. 1

    Hey Adam!

    First of all congratulations!
    I have a few questions if you don't mind.

    How did you make money if your software is free? I saw your answer to a similar question but it's not obvious from the website how you charge businesses.

    How did you acquire your first customers? Do you have a large following on Twitter or somewhere else?

    Can you share your ARR before you sold the company?

    Thanks!

  13. 1

    Congratulations @adammenges That's amazing. I have 2 questions.

    1. How many iterations did it take for you guys to fit product-market fit?
    2. How did you make it through the dark days in those 2 years when things weren't so bright?
  14. 1

    Nothing much to add, but just congrats :) Inspiring story!

  15. 1

    Nothing much to add, just that I love the design & simplicity of lobe.ai, and huge congratulations!

  16. 1

    Pretty impressive! I'm gonna build my AI startup soon and this is a true inspiration.

  17. 1

    How long did it take you to build MVP? I guess a better question is how many hours?

    1. 1

      Oh I didn't skim it slow enough. 2 years.

  18. 1

    Nice app! I like products that make you imagine what you can do with them. There is a book called badass design and this is the first product for a while that feels really badass.

    While I am sure the technical side is very hard it is presented beautifully and simply, and I feel tempted to give it a go for fun.

    A complementary product might be some tiny “spy” cams you can use around the home or a business which provide the input. For example you might set one up on a shop to count foot traffic using the ai.

  19. 1

    Wow it is true nich ...
    And also it is desktop app ... which is rare in out Saas hipe times .
    Smart way to take advantege of personal computer power .
    Can you share your tech stack on the desktop and what part of the process is done on the server side ?

  20. 1

    This is sweet! Congrats, Adam!

    My question is: how does Microsoft plan to use this in their company? (if they told you)

  21. 1

    Congrats Adam... you have officially dispelled the myth of if you build will they come!
    Hats of very nicely built interface and great demo video... is that created by you or you co-founder?
    How did you plan out the marketing so folks could get to you? Love the clean uncluttered look of both your (should I say Microsofts now) website. They should take a cue from you on designing windows!

  22. 1

    Woah 8 figures - 3 founders - that's a lot of money each - that must be feeling amazing considering it only took 2 years!

    Have you ever thought of giving up? When was the time you were down bad the most?

  23. 1

    What was your tech stack?

  24. 1

    Was this your first entrepreneurial project? How did you meet your 2 friends was it school/work/through hobbys ect.

    1. 1

      Was this your first entrepreneurial project?
      Yep!

      How did you meet your 2 friends was it school/work/through hobbys ect.
      We were all working on the problem through side projects. In general I’d consider that the best way to meet people. It’s a clear indication they’re interested in the same things. After we met we had about a six month period where we were loosely working with each other to make sure we all got along and liked each other before founding the company.

Trending on Indie Hackers
29 days left before 2022 🔥 What do you want to finish & accomplish before the end of the year? 23 comments Rejected from YC 20 comments People found our landing page confusing. 9 comments Bootstrapping a SaaS that uses AI to explain code in plain English 8 comments What's the hardest thing about building a profitable SAAS? 6 comments Can you roast my website, please 🙏 3 comments