Lessons Learned While Running an Agency

Earlier this year, while we were shutting down Impact Upgrade, I started jotting down some notes about lessons learned and things I would do differently. I intended to make it more of a full write-up. But I’m instead being lazy and just throwing it out in the void in a raw format. If anyone wants more details, I’d be happy to lean in more…

• Even though it looks great on paper (recurring revenue!), don’t try to fully productize your services. The team needed to run a product offering is a different company altogether. And productized consulting services rarely work without the actual consulting alongside it. Build tools to make your services efficient and somewhat more scalable, rather than try to remove the need for your personal expertise.
• We built up a huge library that was our integration and workflow engine, as we served organizations. On the one hand, it became a secret weapon and allowed us to rapidly repeat results. But on the other hand, it was too much optimization and library building early on. We should have spent less time on one library with major abstractions, and more on niche/standalone tools. That was before gen AI was much of a thing, so it’s even more true now.
• Let tough clients go sooner, protecting the team and reputation. Releasing a client is less detrimental to your reputation than you’d think, but holding onto them and letting things sour further is way more detrimental than you’d think.
• Similarly, release your good clients (even your favorite/best ones) once goals are met and they no longer need your specialty. Don’t let lengthen the contract and continue to find ways to be helpful because you like them and the mission. You’ll wind up in justify-every-invoice territory and sour the relationship. Consulting is meant to be a finite engagement and not permanent support.
• Perhaps the one that hurt us the most: Don’t chase corporate partnerships. We spent a TON of time on partnerships with big companies, positioned as integration and consulting partners for their products. It looks good on paper, but rarely produces the fruit you expect.