April 25, 2019 · work

Chapter 12: Practical advice is often the advice you're most likely to ignore

Almost 9 years ago I helped start a company called We&Co. We were an Atlanta based startup focused on the hospitality industry (a far cry from what we're doing today but that's a blog for another day). At the time, Scoutmob was the ATL startup focused on hospitality. We reached out to the founders, Mike Tavani and Davíd Payne and asked if we could pick their brains a bit. We honestly didn't expect a response because, you know, they were real cool guys working on real cool stuff and we were just this dumb little company. Much to my surprise, they said "sure, come on over".

We met with them after we'd been a company only a few months. We had an app that had just landed in the app store and not much else. We spent about an hour with them getting some real damn good advice I'm sure. Often when you meet with more seasoned founders you get really lofty advice that's hard to wrap your arms around. "Build something people love". Well no shit. Thanks. Or "make it look awesome". Great, thanks for the hot tip.

But in this meeting, one piece of advice really practical advice stuck out to me:

"Automate nothing right now. You'll be tempted to build this gnarly backend admin thing that automates a bunch of stuff. Don't do it. You have no idea what you're actually building. Therefore, you don't know what to automate".

I recall nodding my head and said "ok. thanks". We left their offices on the westside of Atlanta and went back to our dumpy little garage office. And what did we do LITERALLY THAT DAY?

We started building a bitchin' backend admin. We automated stuff. We built in all these analytics. It was neat stuff man. Our customers were gonna love it. It took us about 3 months if I recall. So that was 3 months of building admin stuff and not consumer facing stuff. Time well spent! Suck it Scoutmob!

Except we had no customers. And they didn't love it. Because what we had built....they didn't need. Or want. We wasted 3 months on something we had been specifically told NOT to do by people who knew a lot more than we did. Such is life I suppose (and such is dispensing advice to new founders; prepare for them to ignore it all).

But it's one of those things that really stuck with me. So much so in fact that I spent the last 20 minutes "cleaning" some data for a beta product of ours that helps healthcare organizations track how much medicine is distributed in remote parts of the world to treat a Neglected Tropical Disease called Schistosomiasis (I totally spelled that without having to look it up!).

What it looks like to clean data captured via SMS

The data is collected with SMS because it's the only reliable transport protocol that will work on any kind of phone virtually anywhere in the world. However, as you can imagine, the data comes in unstructured. Given that this product of ours is still in beta, it would be tempted to automate the cleaning of the data. That is, write a bunch of code that can extract the values we want from the mountain of text that can (and does) come in via SMS. This is year THREE of the client using this part of our platform and to this point, we've only automated about 75% of the data cleaning. The other 25% is done by hand. By our team and yours truly. Automating the remaining 25% would take weeks and for what – something we can do by hand in about an hour? On top of that, seeing the actual data helps us understand what's actually happening in the field (going to the field is also helpful).

Working with our partners to see how our SMS platform can help them better collect and track data in the field. Emily Tunggala on our team is either meditating or sleeping while standing. Not sure.

Will we automate the rest of this? Right now we're only doing this for 1 client in 1 country so the answer is....no. But if we scale this beyond a single client in a single country? Maybe :)

Many thanks to Michael and David for being patient enough to wait 9 years for me to take their advice.