Pilot isn't "more bookkeeping software", it's "hire this company so that you don't need to worry about getting into the weeds in QuickBooks Online every month, because you have a person who will do all your work in QuickBooks Online for you (assisted by our software)".

It's not that bookkeeping is rocket science or totally unlearnable (you can definitely learn it), the question is really whether it's actually the best use of your time to master it (when you could instead be focusing on running your business).

Thanks for all your replies in this thread, @wdaher!

Can you share how you are approaching growing / scaling the human side of your bookkeeping army? E.g. the dedicated account managers — when I shared your site with a CPA friend, he asked if you have outsourced teams in India doing the categorizing.

Great question, and definitely one of the operational complexities of the business.

Our thesis is: by relying heavily on software from day one, we can actually do it with a small, super-highly-skilled team (rather than a giant army of folks.)

If the task can really be done by an arbitrary person somewhere without context on the business itself, I suspect we can execute that same task ourselves in software (only with much higher consistency and much lower cost). Transaction categorization is a great example of this, actually.

I think the alternative—scaling by just hiring a ton of people—is compelling because it's short-term easy but long-term hard. By analogy to a software system: if you scale with an army of people, you're building a giant distributed system of super-unreliable hardware (the error rate of people is pretty high!), and you can't reprogram it by running "git push".

I'll venture to make a stronger statement: if you care at all about consistency of quality, I think you have to do this mostly in software -- because otherwise the quality of your service is going to be entirely the quality of the individual bookkeeper you use, which is fundamentally going to be pretty variable.