Note: This was originally tweeted as a far-too-long-should-have-been-a-blog-post thread on Twitter. It was the first thread of several (second, third, fourth) threads leading up to the launch of Otherboard.com. Very minor edits have been made.
There’s almost certainly a fancy German word for the feeling when you are days away from launching your first SaaS product that you thought would take 4-6 months to build and it took almost two years 🤣
This is about to be a heck of a 🧵. Allow me to note a couple of things before we continue:
- This is very much a stream of consciousness. I’m going all the way back, remembering everything along the way, and bringing you through my experience.
- Yes, I know this should be a blog post. It will be (oh look, now it is 😄). I hope twitter threads keep working for a while (plz and thanks @elonmusk.)
Anyway, the month was December. The year, 2020. The day, the 19th.
A food blogger client reaches out and says “what do you use for a CMS?”
I say, “WordPress, dummy”. Spoiler alert: it was I who was the dummy! They continue to describe what they mean: integrated data sources with blog content, phase management and editorial calendar workflow, team management, and ROI insights.
Holy crap, we’ve been misunderstanding what “content management” actually means to people y’all. at least when there’s a real revenue-generating business behind that content.
I share the very genesis of this whole SaaS idea to illustrate one important point: every good thing starts with listening. Really, really listening. This one client represented a segment of our clients. We started reaching out to them. They all validated the same pain points.
This was a real issue that hadn’t been met in a great way yet. Now, anytime I hear folks mention wanting to start something, but not sure how to assess product market fit, what to build, etc…just like…go to the market and listen.
The exact thing they ask for? Might or might not be what you build – but you’ll learn a ton by listening. So. “Listening” started with our clients that were the market for this.
Then it grew to others in the market who weren’t our clients.
But then “Listening” began to look a little different. I knew enough to know that I knew NOTHING about building a SaaS. I began listening – in a fresh light – to voices I already trusted on topics I hadn’t previously given attention to.
Folks like @chrislema, folks like @asmartbear, and others.
People several orders of magnitude smarter than me. Who were GENIUSES at this thing where I was very dumb. I tried to list my “ability to call @chrislema about stuff” as an asset on my balance sheet, but my CPA didn’t get it.
But still – and this is maybe early January/February of 2021 now – reaching out to folks like him to validate the idea was a massive help as I dove in. But what does diving in look like when you don’t know how to swim?
Well, you guessed it – more listening!
This time, diving deeply into worlds I hadn’t been a part of – startup accelerators, @MicroConf groups, food/DIY/Craft blogger Facebook groups, etc. I spent months trying to do two things:
- Recognize how vast the chasm really was between what I thought I didn’t know, and what I really didn’t know.
- Use the information gathered, leaning on smarter folks than myself, and try to bridge that chasm.
- Profit? (jk)
But legitimately, it was (and spoiler: still is!) so, so, so much listening and learning. This iteration of listening led me to one of my earliest technical (and to date, best) decisions. Using @laravelphp to build the SaaS. I decided to take roughly the entire month of February 2021 to do as deep a dive as one could do into Laravel.
Poring over the documentation, joining all the Slacks, finding and following key leaders, researching canonical resources, and finding small ways to contribute. Being open source and PHP – there’s a lot of lovely overlap – even among community members! – between WordPress and Laravel. Both communities learn (and still have much to learn) from each other.
Comparing the two may be a thought for another day – but for now – love them both. After spending the month building, learning, and prototyping with Laravel – had my initial estimates around a 3-4 month build.
Oh sweet, sweet, dumb, younger Justin. How naive and beautiful you were. So I tried. I really really tried.
We had great clients and our team was crushing it at @zaowebdev – so I could afford to step back for a few months and just deep-dive and try.
For someone who had no idea what they were really doing – I did ok! talked with some geniuses: Had some amazing design work from @shaybocks and some brilliant head-starts from @jasontheadams on the best path forward, and I actually had a really cool, functional prototype that we let some beta testers use in June/July of 2021. BUT THE STORMS WERE A’ BREWING
I knew going into it that it was likely v1 of a few versions of the app. Nothing was precious. But the beta testing group (shout out to you all, you know who you are!) made it clear to me, I had too much technical debt, and scaling issues would come soon.
I needed reinforcements. so this is inside baseball a bit, but you know what the hardest part of running an agency is? Tough clients, deadlines, trimmed budgets, internal team drama, logistics, quality control, communication?
Nah. for me, anyway? Hiring. HARD to do well. In WordPress, which I know well! So now imagine “dummy-knows-very-little-Justin”, hiring killer Laravel developers to bring this vision to life in a scalable, sustainable way.
Again, with the spoiler alerts(!): it took a few tries. More on that to come.
At this point – it’s July 2021. I’m burnt out. Things were getting busier at @zaowebdev – and we were growing and hiring more there as well – and managing that on top of this was starting to kill me. So I took August mostly off from dev work and focused on hiring. Interviewed around 100 Laravel developers. I have…many, many thoughts on the state of DX/developer experience levels in WP vs. Laravel.
But they are very, very different.
But also? kind of the same. In that, you can usually get a sense pretty quickly of who really knows what they’re doing and who’s sort of faking it. By the beginning of September, we had 3-4 folks shortlisted and working together to assess what I had built, review business logic, and make assessments as to whether or not what I built was usable or not.
Good news, it was!
Spoiler alert: it was not
So I head back over to Zao-land – we had separate, significant initiatives going with both @Pagely and @GoDaddy (pre-acquisition, lol) – and it took most of our entire team’s time, mine included.
But hey, the new SaaS team was killing it, using my code, let’s go! So aside from weekly check-ins and the odd review here and there – they were essentially on auto-pilot in September and October 2021.
It was 100% my bad for taking my eye off the ball with them. Looking back, it wasn’t their fault at all, it was mine. I don’t know, in that season, that I could or would have changed anything…but it set us back a few months.
Still needed reinforcements…just better ones. So we’re solidly at “12 months ago” now – 3-4 months past my estimated launch date.
I start re-scouring “Laravel Twitter” and “Laravel Slack” – places I went to learn before, and I now went to hire. Always be learning. Found some lovely, brilliant developers. They were way smarter than me, and way, way smarter than the other folks we had working on it.
They took two looks at my code and said, “Hey nice try! scrapping it all.”
I cried a little. And we scrapped it all. It was a fantastic decision. Lesson learned: sunk cost fallacy is real! But the “fallacy” part of that is key.
Don’t lose yourself on sunk costs. Everything is an experiment. Experiments teach us a lot. “Winning on your first try” isn’t the right definition of success. So we contract these lovely, brilliant Laravel folks (you know who you are, thank you!) between November 2021 and February 2022 to rebuild the entire app.
From the ground up. Burn every assumption to the ground, nothing is sacred. And they did! And it was beautiful to watch. I’m profoundly aware of how ridiculously privileged my life is that I can pay a fantastic team of devs to help me build something amazing – and use that as a 4 month learning opportunity.
It is and was exactly that, a privilege. But as they say, every new beginning comes from some other beginning’s end. Contract was up. Estimates were “1-2 months of work left”. Re-engagement didn’t pan out. So with a solid new foundation in February of 2022, what do I do – try and hire again or try it on my own?
At this point, I was far more confident in my ability to hire good developers than I was to be a good Laravel developer 🤣.
Back to the hiring board!
Between February and the end of March, I reviewed 350+ applications. Let me tell you, Justin-from-February-and-March-of-this-year made an excellent decision. I was a bit slower to hire this time than the previous summer. Less desperate. More focused on the long-game and folks who would be around awhile.
It paid off, y’all. Ended up sifting through 300 applications from @laraveljobs (shoutout!) trialed about 15 developers from there. Found our first full-time hire from this group (he’s amazing and still our lead engineer). then we had ~50 or so applications from various networks – Slacks, Twitters, etc. Our next hire was from @JackEllis‘s brilliant Serverless Laravel Slack.
Maybe a separate thread in due time on our tech stack (AWS, Vapor, SQS, SingleStore before too long, etc.). So we’re now in June 2022 – two incredible developers with us – cranking away on building out the SaaS vision.
At this point, I’m working with them to assess whether or not that “1-2 months of work left” was real or not. so as it turns out, estimating complex software development timelines is hard. like, way hard.
We were (again with the spoilers!) not 1-2 months away. But how far away? By roughly the end of July, we had a pretty good idea that we were 3-4 months out. Mid-August rolls around, we add another developer to the mix, plus our PM is splitting time between our SaaS and Zao. Five of us, an end in sight, let’s go! Something funny happens when a project goes from the abstract “it’ll be done when it’s done!” to “here’s a deadline we’re committing to”.
Again, German friends, you must have a word for this? Time both slows down and speeds up rapidly. Bugs multiply and reproduce. It’s crazy.
A couple of asides that I don’t want to skim over because they were HUGE wins:
- In the middle of all of this – Pinterest opens up their API to us. MASSIVE win.
- Our deep integration with Google, Facebook, Instagram, and all the others – all of which have long approval processes? ALL APPROVED!
- WordPress plugin? Approved! Now it’s September of 2021.
Notice all I’ve really talked about up to this point is development? As it turns out, building a SaaS is a lot more than that. In this season (~summer/fall 2022) – we had Gantt chart breakdowns with over 15-20 people on them at times.
Conversion rate optimizers
Designers (shout out to @VinThomas and @shaybocks!)
And that’s just some of the non-engineering folks. I guess what I’m saying is that there are definitely people who can do all of these things on their own – or at least appear to.
I’m not one of those people. I’ve needed, and been so thankful for, a TON of help along the way.
Remember – I have no idea what I’m doing. Lol, speaking of not knowing what I’m doing.
Many, many, many brilliant people – from validating product market fit, to writing excellent copy, to designing intuitive interfaces, to selling and marketing it all brilliantly, to building it scalably…so many others are doing work behind the scenes. Some day, I’ll have the opportunity to highlight each one of them specifically and show you why I think they’re so special and amazing.
But we’re now, let’s say, 6 weeks ago. Solidly into late September. We have a clear vision of what’s left and where to go. At some point at a team meeting, we say “ok – here’s what we’re launching with, this is what we have to nail. Everything else, we cut”. Those conversations are hard and can hurt.
Dut delay is rarely denial (true in engineering, true in parenting…generally true!)
We’re deep (I mean…deeeeeep) in QA, debugging random SQS timeouts, API rate limits, UX issues. Dealing with small team resourcing stuff (vacations here, maternity leave there, holidays, etc.) And still, everyone is crushing it. Eyes are on the prize. We take all of October and focus deeply on any of those core launch items remaining, fixing other bugs along the way.
We fix a lot of big bugs, and find some new littles along the way.
Currently, our Github tracker has 33 open bugs and 33 closed for Launch. I like that. At this point – everything in the app is functional and complete. We’re in bug-squash mode for the next little bit here, but we’re actually gonna launch the thing! At different points and in different demos – beta testers have told us what a game changer this will be for them and how powerful it is.
I really think it will be neat for the market it serves.
With this SaaS (and Zao, and everything): my mission is margin.
“Hey Justin, dummy, what does my mission is margin even mean?”
I’m glad you asked.
In the context of this app: it’ll give people back hours of their day (and dollars in their banks). Getting time and money back in your life is really powerful. And when we can leverage tools (like businesses) to give people more…life…in their life? That’s what it’s all about, for me anyway. So those are the stories I’m stoked about.
Not the stories about “what a cool app we built!”, Nah. I’m excited for the stories of our clients and people like them, coming to us saying,
“Man, you gave me 45 minutes back in my day. That’s not a lot….but it is a lot, ya know?” or “Man, I never know what the ROI on my blog posts actually were. Now that I do, I’m making more money!”
That’s what gets me going, y’all. Anyway – that brings you up to speed! I’ve got my Pre-Launch Checklist in hand. All the folks on all the teams are crushing it.
Oh yeah! it’s called @OtherboardApp 🤣
We launched it on November 20th.
More to come 😎