Day 400: free freelancing

product

today at my co-working space, we had a special chinese new year lunch where we all sat around big tables and ate some amazing vegan mediterranean food

i was looking forward to this event because there actually aren’t too many opportunities to meet people socially in co-working spaces. i’ve gone up and introduced myself to some people who i see really frequently, but if i’m not sitting close to them on a regular basis, it’s unlikely that we’ll ever strike up a conversation.

so we’re sitting at these long tables chatting about how we all ended up in taiwan, what we’re doing here, yadda yadda, the usual. someone asked me what i’m working on so i talked about browserflow and how it’s the beez kneez

one of the guys i had just met said that he’d played around with browser automation, and that he had tried using katalon recorder in the past to automate some stuff on amazon but it hadn’t worked.

i got really freaking stoked because it’s not every day that you hear someone talking about some random chrome extension that you found while doing market research months ago on everything related to browser automation. (katalon is not obscure by any means especially if you’ve looked into automated testing, but it’s not something you expect to hear at a casual lunch)

i told him i’d be down to help him out because at this point, i’m confident enough in browserflow’s stability and versatility to handle the majority of use cases people have

today afternoon, we sat down and i got to learn more than i’ll ever need about amazon sellers and the marketplace for amazon businesses. he told me about how, as a broker, he uses a virtual assistant in the philippines to do all of his lead generation to find amazon sellers whose businesses he’s interested in selling to interested buyers.

at a high level, he has his VA look through a bunch of amazon listings, get information about the seller, determine if the seller meet his criteria, and if so, add them to a spreadsheet.

once they’re in the spreadsheet, the goal is to get in touch with them, so the VA then uses a bunch of heuristics to go from the storefront or brand’s name to a phone number / email / social media accounts, whatever.

when i had him explain the step-by-step process that he has the VA go through, i realized that this was perfect for browserflow. it’s a recurring task that provides business value, aka the kind of people would be willing to pay for.

before the end of the day, i had working flows for both the lead generation and contact information searching heuristics. it all just worked

i realized today that i’ve been sitting on a goldmine without realizing it: i’m at a coworking space in one of the few places in the world where i can sit next to a stranger and have them tell me all about their business processes and what kinds of automations they’re looking for. i’ve been sitting in the corner coding all by myself when i could’ve been working on real use cases and directly learning what kinds of solutions people need!

i’ve worked with some people over video chat on their flows, but there’s nothing like sitting side-by-side for hours where you can ask clarifying questions in real-time, come up with heuristics together, and then have them try out a working flow within minutes or hours

i got all excited about this because just like i did back in austin, i’m going to set up a big sign with something like “I’LL HELP YOU AUTOMATE YOUR WORKFLOW” at my coworking space and do a lot of sessions where i help people using browserflow. my priority right now isn’t to sell — it’s to learn more about the kinds of problems people have and whether browserflow can solve them

i’m basically doing free freelancing, but i benefit as well because i get to see where browserflow is still weak. for example, here are some things i noted while working on today’s flows:

  • things can get pretty unwieldy with long workflows. browserflow allows composing flows together, but the support is pretty rudimentary and there’s a lot of room for improvement when it comes to usability. i thought about going deep into this, but flow composability seems like the kind of thing that can wait until after i have paying customers who are happy with the core functionality
  • sharing flows is tricky. when you import a flow, you create a copy of it (think of cloning in github), but unlike github, there isn’t an easy way to get back to sync with the original except for deleting yours and reimporting the updated flow.
  • browserflow currently doesn’t have robust support for recovering from webpage errors (e.g. in one of the flows, the website we used sometimes returned an error for no reason) which makes it brittle on flaky sites
  • there were many times when i thought “oh this trick i’m using is not going to be obvious to someone with little web programming experience”. trying to think of how i can lower the bar as much as possible, but i think there will always be some tricky bits

anyway, there are a bunch of things like this that only become clear from real world usage. too bad i have the worst timing: the co-working space is closed starting tomorrow due to chinese new years, so i’m going to have to wait until next week to unleash my plan to get everyone here using browserflow

let’s goooooooo