Editor’s Notes: In this post, the CEO talks about i) Profitability ii) Pricing iii) SalesOps iv) Management v) Hiring vi) Partner Pay Model vii) Product viii)Agents ix) Operation and x) Data, Analytics & Automation.

The Hindsight Exercise.

We’re on track for profitability by the end of this year. But it took me 5 years and $5M in capital to get there. If I could time-travel, so I had the gift of hindsight, knowing what I know now, I should, in theory, be able to achieve the same outcome, in at least half the time, with at least half the money.

There’s two ways to read this list. Either as our Top 10 Biggest Wastes Of Time And Money, or as a Cheat Sheet From The Future, written to my past self.

1. Profitability. Ironically, I always thought I was under-capitalized. In retrospect, I was over-capitalized. The problem is that I wasted most of the money focusing on the wrong metrics, trying to achieve VC rocketship growth. If profitability had been the unalterable, unquestionable goal from day one, then all incentives would flowed downstream from that. Profit discipline forces strong accounting, income statement awareness, margin sensitivity, ROI-weighted decision-making, and overall efficiency. For example… Having a big R&D budget and a tiny runway probably results in waste. Whereas having a small R&D budget and a huge runway probably results in innovation.
2. Pricing. It took us too long to arrive at our current pricing model. We have three simple plans. They are minimums, but clients can spend as much as they want with an automatic top-up mechanism. They are pegged against hourly rates, but within the first fifty hours of running a custom process, we switch to results based pricing. Enterprise clients get automation discounts. Our current pricing is not our final pricing; it will continue to evolve, and we have an exciting vision for it. But there were years in which our pricing didn’t change, so we weren’t evolving fast enough.
3. SalesOps. It took us years to find a distribution channel that worked. We wasted a lot of money on traditional sales and marketing approaches. It took us too long to invent a unique, home-grown SalesOps machine that could systematically target and reach out to customers, so we could replicate successful use cases and test targeting hypotheses. If we built this Unlimited Meetings Machine sooner, it would have accellerated our growth trajectory.
4. Management. I became a better manager over time, but more importantly, I hired better managers in Marshall, Scott and others. Historically it took us too long to let go of partners who weren’t a fit. Also it took us too long to design an org chart and operating systems that would minimize silo’ing and maximize performance.
5. Hiring. I tolerated a weak Partner hiring function for too long. When you’re desperate to achieve your objectives and you only have one candidate in front of you, you may make the hire. When you’re desperate to achieve your objectives but a team member is underperforming, you may threaten them but you may not fire them. And that makes your threats empty threats. In both situations, you are effectively being held hostage by having a weak hiring function. When you have a strong hiring function, you make hiring, management and firing decisions from a position of strength.
6. Partner Pay Model. Related to hiring, it took us many iterations to arrive at our current Partner Pay Model, which ties compensation to Net Income — the bottom line. Every improvement in our PPM has improved our hiring process, filtered out bad fit candidates, and aligned existing partners with clearer, more motivating and more aligned incentives. The more clear Invisible becomes on its own culture, values, vision, incentives and business overall, the fewer candidates are going to be a good fit, and a good PPM allows them to self-select out.
7. Product. Sequencing is hard. I wasted a year trying to solve a very difficult information architecture problem that we probably won’t have to solve for another two years. But the right lesson isn’t to avoid all hard problems: if I’d have done that, Invisible wouldn’t exist. Hard problems are opportunities. The right lesson is getting the sequencing right and understanding how they relate to each other. We also wasted a year building our first version on top of Slack instead of in our own interface, with the rationale that it would save us money on developing our own UI and we wanted to meet users where they were at. The right lesson? Francis isn’t a great Product Manager. If I’d gotten hiring right, I wouldn’t have been making decisions at that level. Another sequencing mistake: we wasted a lot of time on process architecture before solving workflow management and labor coordination problems. It turned out that a competent Ops team could provide the last-mile guidance around the work product itself. But managing the overall delivery workflow was what allowed the Digital Assembly Line to support its first $1M ARR in revenue. Only now is Process Builder becoming the backbone of our automation strategy.
8. Agents. It took us a couple years before we hired our first agent. A shockingly long time. If we had done that sooner and focused on just being a low-tech agency making a simple arbitrage, we might have had a better evolutionary starting point, and a better lab environment for product development.
9. Operations-First Culture. Tech culture is Product-First culture. At Invisible, we’ve evolved an Operations-First Culture: we’ve noticed that the closer everyone is to the clients and the closer everyone is to client work, the better our products and marketing become. Operations cultures value density over specialization. The traditional logic is the more people you have, the more specialized they can be, therefore the more productive they can be: marketers focus on marketing, engineers focus on engineering, etc. In this paradoxical logic, the fewer people you have, the more aware they are of the whole, the closer they are to clients, therefore the more collaborative and effective they can be: marketers aren’t on the marketing team, engineers aren’t on the engineering team, everyone is on the Client Team.
10. Data, Analytics & Automation. It wasn’t until a year ago that we built a data warehouse and started building analytics dashboards on Metabase. And it wasn’t until a year ago that we set up agent monitoring software to collect all of the data from their machines. And it wasn’t until six months ago that we started focusing on ad-hoc automations. Although systematically automating structured data is The Holy Grail because the automations are re-useable, even ad-hoc automations usually make good ROI sense.

Now, what’s the point? Was this exercise valuable?

Hindsight is an unfair advantage. Nobody knows which clients, hires and features work out in advance, and which ones don’t. That would be God Mode. So, of course someone with a hindsight advantage would outperform. That’s not the point of this exercise.

This exercise is about observing how our approach to the fundamentals shifted over time. We’ve got a different approach to strategy, sequencing and prioritization, org design and incentives, pricing and positioning.

A few abstract thoughts, in parting:
— Evolution is powerful but it responds better to time than to money. Profitability, strong pricing, aligned incentives, capital efficiency, financial discipline, ROI-weighted decision-making… all of these things buy you time. One day you’ll wake up, and you’ll realize that finally, time and evolution are on your side, and you’re no longer in a race to survive.
—Entrepreneurship is a multi-front war. Once you have a virtuous iteration cycle in an area, compounding takes over. But getting virtuous cycles going in all areas at the same time is what results in victory. This results in a game of wack-a-mole. Especially as all areas interact in complex ways, and sequencing is hard to know in advance.
—Do not trust experts. Do not follow formulas. Read the books and blog posts that promulgate the dogmas, but then do your own first principles thinking based on your business. When you break from consensus and you are right, you will be rewarded massively. But until you are right, you must have the courage of your convictions, because everyone is going to tell you you are wrong. Meanwhile, you can’t do everything yourself and you won’t be right about everything, so you must question yourself — just not in the same way that everyone else is questioning you. Hire other first-principles thinker/doers that you respect. They are very rare rare and hard to find. Once you have a Stem Cell like that operating in each area of the business, virtuous evolutionary cycles take over and compounding begins… After that it’s just a survival game, as it’s a matter of time for the game to shift in your favor.

Someone used to tell me: “if you’re not embarassed by who you were a year ago, you’re not growing fast enough.” The Hindsight Exercise is an embarassing one to run, because when I do it honestly, I see a list of my biggest mistakes. I wonder why my education was so expensive and so long. Am I just that thick? Maybe I’m not cut out to be an entrepreneur after all, or I’m just not that good at it? Will all of these hard lessons pay off someday? All of these self-defeating questions arise.

But although I can spend an hour or so writing these lessons down, and turning them into a list… it’s mostly for my benefit. I wonder how much a reader can really gain from reading these. Experiential knowledge is far more complete than theoretical abstraction.

I’m a theoretically-oriented person. That’s one of my strengths, but also one of my weaknesses. For example, I look at our client base now, and I wonder, in hindsight, what books could I have read — on industrial history, on operations, etc. — that would have given me the clues I needed to target these clients sooner? But it may be that any theoretical approach to solving this problem would underperform an experiential approach.

Experiential approaches seem stupider, but they have evolution on their side, and evolution is smart. Our SalesOps Machine started out like an amoeba looking for sugar, but increasingly it’s becoming algorithmic, and the more plays we feed it, the harder it is to reverse-engineer exactly why it is targeting this person or that person. My intuition, its intelligence, and our business performance co-evolve.

This is a question I suspect I will never have the answer to, but I will always keep asking… A Koan of sorts. Let’s call it The Innovator’s Koan:

HOW DO YOU UNEARTH WHAT WILL BECOME OBVIOUS-IN-RETROSPECT? HOW DO YOU SCALABLY DO THAT?