I was the co-founder of a 50 person marketing agency. Left to learn how to code, and my EdTech product just had its first $1k+ week. AMA
I co-founded Ladder.io, a growth marketing agency, and grew it to 50 people over 6 years. I was in charge of operations, so spent most of my time on training the team, solving problems for key clients, and automating the work with "no code" tools and in-house tech we built.
I got more and more interested in the product side and tired with agency life, so I left my co-founder in charge, and parted ways to spend the year learning to code. I built a few different things and eventually went all in on one idea, simulator-based courses to teach marketers how to be more data-driven.
I work 3 days a week on Vexpower.com with my co-founder who joined after I had built the initial prototype, and consult on marketing attribution the other 2 days a week to pay my expenses. We just had our first $1k+ week so we're hoping next year we can do 4 and eventually 5 days a week, and scale it.
Ask me anything!
Who says bootstrapped means small?
QR Feedback - Marketing Ideas?
Conquering On-Call Burnout: Essential Strategies for Tech Teams
Hey Mike! As someone with marketing expertise who now understands devs, what is the 80-20 of "getting good at marketing" which most dev IndieHackers would do well to learn?
My advice to devs is to stop trying to "get good at marketing" - you can't treat it like an optimization problem until you get to scale. Instead try and sell to one person at a time, until you have a good idea of who they are and what language to use to get them excited. Once you have that, you just need to put that language in places these types of people hang out.
Just wanted to say that its awesome! I left AdTech world and all the expertise behind to do something similar because I didn't see it evolving and expanding. Coming from a growth background too there is definitely some cool stuff you can utilise on your own products but learning to code is gruesome sometimes particularly if you are doing it part-time.
I always say - most resources out there teach you how to get a job, not to get dangerous with Next / Tailwind / DB kind of projects - perhaps also a tip for quite an untapped market.
All the best!
yes everything assumes you want to get a job, but frankly as a self-taught dev you can make a lot more money and get a lot more recognition earlier on by building your own stuff or working with clients
Ed tech is the future. Really interesting product, I'm going to give it a go. Thanks Mike!
awesome thanks for the encouragement!
Hi ! what an amazing journey you're on, very inspiring for a young entrepreneur as me.
Just finished my engeneering school and I want to challenge myself to build a marketing agency as co-funder who will ensure the tech part.
So my question is linked to the fact that I saw your client portfolio and it's impressive so how you make the difference what i mean by that is what do you reach client that will make the difference in your business? Fact is that selling website isn't a sustainable way to grow our business.
We got most of our clients initially by referral, basically we went to meetups and told people what we did, had coffees with them after, gave advice to impress them over coffee and they would then refer us to clients. Eventually we started blogging and that became 60% of our leads, and PPC was another 10%.
Hey Mike, I've always been a huge fan of Ladder.io content—particularly the technical depth. Congrats on your traction with Vexpower. The training simulator is a great concept. Will definitely check out the course.
awesome thanks - all my experiences at ladder are being channelled directly into the stories the simulator-based courses are telling
Super inspiring, Michael! As someone who works in the marketing world and wants to learn to code, I’m curious to know how long it took you to make the transition? Did you have any technical background at all?
I studied Economics in university but it had no real Math requirement (it was a glorified business management degree). The limit of my technical skills was how to do pivot tables and vlookups in Excel for the first few years.
Then I took the One Month Rails course (Mattan Griffel) like 7 years ago and caught the coding bug. It was right about the time Growth Hacking was taking off, which made coding to grow seem a lot cooler than just running ads. I was flying to San Francisco every month for work, and I was such a huge Silicon Valley fanboy.
I did tutorials evenings and weekends for about 5 years then started coding 20 hours a week when I left Ladder last year, 40 hours a week this year. I added it up and I've done about 2 and 1/4 years full time equivalent, only 9 months of that was over the first 5 years (evenings and weekends don't add up to as much as you think). I had just gotten to the point of being able to build basic stuff when I quit Ladder.
So long story short, if you quit your job and started coding full time today, you'd probably be able to build something decent within 9 months. Within 2 years you could be getting paid a very good salary to do it full time (or be full time on your own thing if it works out). If you do it part time it'll take a decade.
Thanks for the super thorough answer. This is exactly what I was looking for! I'm in a similar boat to where you were seven years ago — business degree, strong Excel skills, but nothing technical.
This is really inspiring and I totally admire you taking the leap! Best of luck.
For context, I've launched Outdone (https://www.outdone.io) with another non-technical co-founder. So we've outsourced the technical build to a team of freelancers, but this is (obviously) slowing down our ability to iterate and test. Of course, I wouldn't be able to build enough tech skills in the near future, but I'm thinking a baseline in Python would allow me to troubleshoot some things and speak more intelligently about the product!
Honestly if you can get a product built by someone you might be able to skip learning to code entirely. Though it might be worth learning a bit to better manage devs.
Fair enough. Having it built by something means its more costly and makes it more difficult to iterate on the fly — but it does have it perks.
Nice work on the milestone, Michael! How do you go about building Vexpower? Did you start with market validation/problem you knew already existed? With your new programming skills and existing no-code skills, what route did you take with building out Vexpower? Was that a custom build or built with a no-code solution?
Thanks! I started off doing the classic Indie Hacker thing of trying to launch one startup a month. This one was the most promising: https://www.zirality.com/
However that tech required cookie tracking to work, and when I saw cookies being deleted after 7 days on Safari due to ITP, and further research gave me a heads up on iOS14, which I realized would be a blood bath.
So that made me think "how are we going to do marketing attribution without cookies?". I went down the rabbit hole on Marketing Mix Modeling - talked to a few experts, did a project for free for an old client, and wrote some blog content.
I tried a lot of coding courses and almost joined Lambda School, but in the end it was https://www.dataquest.io/ that got me over the edge. They have an interactive coding terminal like Codecademy but the content is more immediately useful, and I like it being self-paced.
By the way, 2 years later I'm coding 2 days a week to pay my bills while I work on Vexpower! Primarily focused on MMM because that's what I live in every day, though I'd say I'm at the point where I can figure out how to build most things that I dream up.
I started building an MMM tool, and found it was basically impossible to automate (at least for me, I don't have a PHD). That's when I came up with the simulator idea: it was my way of guiding the user to make the right decisions when building MMM, but in a more dynamic environment.
I validated that idea by stringing together a few Typeforms on my Webflow site to make a simulator, and then charged pre-orders with the Typeform Stripe integration. That made me about $1k so I felt like that was enough validation and started 3 days a week as my main focus (this was January 2021). www.vexpower.com/access
Version 1 of the platform kept the Typeform front-end, and I used the Typeform API to programmatically generate unique Typeforms per user based on what they had answered in previous Typeforms! I also built a Flask site to host it all on.
That was super clunky and traction had slown down after a month or so, and everyone I talked to was confused as to whether I was a SaaS tool for MMM or a course teaching MMM. They all said they didn't trust a 'black box' solution and just wanted to learn MMM, so I made it an education product.
That's when I moved to the current idea, which is basically Udemy for Simulator-based courses. I rebuilt the app from the ground up, this time focusing on front-end first before building backend, and was much happier with the result. I used a NextJS React stack, with Tailwind, which really made a difference in terms of iterating faster and it works a lot smoother than my clunky flask app.
It was slow going though because I had to learn React from scratch to build it (and my JS was pretty poor too). I learned how to do React from different developers I knew - paying them consulting fees to pair program with me worked best. I got the prototype working, and then managed to get my co-founder James on board, who is a much better programmer than me (but also had to learn React from scratch, as he was a Python Data Science guy).
It was quite a journey, and I would be able to skip a lot of it if I did it again, now I'm a much stronger coder and have a technical cofounder. I think what I did that was smart was to build it in public, and bet on the right framework. I think it was smart to do the prototype in Typeform, because it taught me a lot about what I was building and how to build it, learning how they do things via their API.
Thanks for your detailed feedback and insights. From a developer perspective it’s interesting to see how newer coders approach problems. Your use of typeform is a great validation method and I believe that nocode tools will continue to enable those new to tech to quickly test their ideas while learning or finding a technical cofounder.
One huge aha moment which I don't think even most developers get yet, is that when I tried to build things from the backend, i.e. a jupyter notebook, I end up with something clunky and unusable. Whereas when I start from the front end, i.e. a mock-up in excalidraw, it comes out a lot cleaner and elegant. It shouldn't make a difference because you need to build both front and back to launch, but I think going front to back rather than the other way, helps you simplify the offer and align it with what users want.
This is great advice. You can usually tell when something was built backend first. I'm rebuilding a system at the moment, and I'm taking the frontend first approach this time, and it's also stopping me from scope creep. Usability is extremely important these days, and functionality can be plugged into pretty much any UI these days.
Super inspiring Michael! Will definitely become a customer!
How do you balance doing two very high-performance activities whilst keeping the quality and fresh mind in each? What are the pros and cons?
I don't feel like I 100% have the balance right, but I think it helped that the way I do marketing was with the engineering mindset, having come from the Growth / Performance Marketing world. I always try and bring things back to first principles, break the problems into smaller chunks, and then just do the work.
It also helps to focus - when I was at my agency my schedule was determined by whatever fires needed fighting, and since I left it's been a battle to stay disciplined and wean myself off the dopamine of doing lots of different things. I've been slowly cutting my focus down and getting rid of consulting gigs that were out of scope.
Now it's pretty narrowly focused and I'm seeing huge gains all of a sudden, because all my consulting is on Marketing Mix Modeling, and therefore everything I learn can become a Saxifrage blog post or Vexpower course, which then gives me content to push people to when they have a question.
The challenge will be staying productive as I expand focus: we're starting to work on non attribution topics for vexpower courses, and I am working with consultants in those areas to keep me fresh and make the courses useful / realistic. To make space for that I've cut my consulting right back and will just have one stable client in 2022.
Congrats Michael!
thanks!
Welcome
Yo Michael, thanks for doing this.
In your experience, what are the distinct "breaking points" of going from 0-50 employees?
At which headcount thresholds are there significant changes?
Thanks
For us the break points were pretty much every time we doubled:
Every time we got to one of these milestones the business fell apart and we had to rebuild it. I'm talking literally change our entire business model and org chart, often turn over 10-20% of staff (its possible we just did a bad job of it).
Often we'd be stuck in the cashflow spiral of "sign a client? need to hire ASAP", "lost a client? Need to fire ASAP". It was only after 40 employees that it started to be more sustainable and manageable. Then at 50+ I was able to step away.
Hi Michael! Happy to hear you're going to focus more on Vexpower in the near future and hope everything goes well there.
I've a rather long question to ask, but I'm hopeful that you're too generous to dismiss it.
At SMEs in developing countries there isn't that much of public, secondary data, and the ability to collect private datasets is pretty limited.
To clarify, I mean that such companies normally can't hire a marketing research service provider, but are still in need for good marketing reporting that encompasses all of their marketing efforts (offline & online).
With such conditions, how one could build a decent measurement system that aims for understanding, valuing past effort and guiding future decisions?
You know I basically did no marketing research over my 10 year career over and above just Googling for things people have already studied. My opinion is that you just need to get out and test it. Research at its best gives you ideas to test, but at its worse it can give you the false impression you're doing something productive while your customers wait to hear from you.
Thanks, Michael!
Hey! Glad to hear how much progress you're making on Vexpower 💪
I'm thinking about spending some time on learning python as a SaaS SEO. Can you help persuade me? :)
What are 2-3 things I might be able to do (either new things, or faster/better) in my SEO work with basic python skills?
Primary uses of Python for SEO:
The bigger benefit in my opinion though is just learning the engineering mindset. Shifting from "I have no idea how that works" to "I bet I could figure it out" is a massive unlock, even if you don't do that much active coding for work.
I wrote more on growth engineering here: https://www.saxifrage.xyz/post/growth-engineering
Thanks a ton!
...so did I convince you?
I'm convinced that I should learn python... Now I just need to figure out how to prioritize / where to get X hours per week from to do it! 😁
block off an hour each morning or a half day on the weekend - that's what worked for me
This comment was deleted 2 years ago.
It changes over time, but for us Facebook Ads clients drove 80% of our profit. I think SEO and content production tends to be more stable but lower paid, Conversion rate optimization is better paid but really hard to do well, and Creative is well paid and less rigorous, but also more subjective with shorter term projects.