Distribution vs. Innovation

This advice is essentially "don't build on top of platforms, because platforms can become competitors".

It's good to point out this risk, but as with all maxims, it can be wrong depending on the details. Many big companies have been built on platforms, after all. In fact, as far as I can tell, one of his examples of "boring" companies that doesn't fall victim to this is Square, and Square is built on top of a platform. It's built on top of the credit card network. But VISA didn't kill them. Why not?

You have to look at the likelihood that the incumbent will enter your market. This depends on the threat/opportunity you pose. It also depends on the technical assets and abilities of the incumbent. It also depends on the opportunity cost the incumbent will incur. It also depends on the strain it will put on other lines of business. If the incumbent has to rewrite much of their business processes to compete, they'll be less likely to do so. You have to look at how open the platform is, and how likely the incumbent is to change it in order to kill you.

This. is. why. marketing. is. so. important.

If people can't find what you're selling, or don't know why they need it, they won't buy it.

So, think about how whatever you build is going to be distributed. Think about leveraging different channels that aren't being used by incumbents. Think about whether your product is a vitamin or an antibiotic.

The author leaves the punchline a little vague, possibly in deference to a16z portfolio interests. My rendition is this.

Value capture for any tech almost requires owning the customer relationship. Customers (non-enterprise anyway) buy must-have solutions, not fancy bolt-ons. Ergo, inventors must sometimes build the mundane parts of an app/platform in order to fully cash in on the fancy new parts.

This need goes against founder/inventor philosophies. It's wasteful and boring. That's what makes it so hard. It's also more difficult to fund building the broad solution.

I think that many of today's AI companies are going to come up against this dilemma soon. Can a tool that analyzes CRM data make as much money as one that handles the whole problem?

(Think of how many times CRM tools have been reincarnated in the last 30 years. Building yet another is not that crazy.)

Once you’d used TiVo to time-shift (up to 1TB with easy upgrade from weaknees) and channel skip (“blip! blip! blip!” for 30 secs) it was intolerable to watch cable live.

TiVo trained a generation of future cord cutters they could catch ‘their’ shows and still have Thursday night free.

If networks had understood this, they’d have embraced ‘apps’ sooner. It’s dangerous to let someone else set your customers’ expectations.

The kernel of the lesson here is an old one but key: own the customer relationship, at least in the beginning.

We wasted a year trusting a huge but disincentivized distribution channel: huge value for the org, but their salespeople didn't care. It ended up ok because we accelerated learning about the rough problem space, learned to own our accounts, and as we go back, have stronger footing ("oh you are taking over OUR account discussions?", "Here is exactly how to sell..").

So in hindsight what is being suggested? That Tivo should have established its own content network and then add the dvr to it? That TrialPay should have created Stripe and then added the trial pay feature?

I think I'm misinterpreting this advice because it sounds to me like the chicken and the egg problem.

This is the fundamental problem we struggle with in Agtech. Large agribusinesses have a near monopoly on distribution and for anyone but the most well capitalized companies (where capital has preceded traction) it’s a challenge to scaleably get your product in the hands of customers. Moreover most VCs are not willing to invest in a company to put in the hard yards. If a16z is backing you, like they backed Granular, you have an opportunity to follow the author’s advice but few first time entrepreneurs will make the cut because a16z is either going to bet on traction (which comes later), the serial entrepreneur with one or more big exits, or the technical dream team. I’d love to see some case studies on how to solve this problem without a lot of upfront capital. Not everything in amenable to an MVP.

This is a good article. Many times "techies" fail to appreciate distribution. I did when younger but as I got older and saw it play out and now appreciate it. A better mouse trap without distribution will fail against a mediocre mouse trap with great distribution.