Silicon Valley has operated startups with the waterfall method for five decades. We have grand visions and act like a pivot is a mistake. That's wrong. Here's a story about how we tried 100 pivot-level ideas at my first company and found 4 or 5 that worked:

Please support by enabling ads. Thanks! :)

When Silicon Valley was about chips or hardware, pivots were expensive. Before AWS and open source, even software pivots were expensive. Now, pivots are free, and pivots can mean 10x bigger opportunities, so you should pivot until you have insane need for your product.

I started three companies, one through IPO (NetGravity), one which is still going strong (Rocketship), and one sold at a fire sale (Zeal). At NetGravity, I validated everything. By Zeal, I thought I knew the truth and would reveal it to the world to loud applause.

Here is the deal. If you think you know what users want without validating, if you think your gut is always right, if you think you can guess how the market will work when you give it something new, you are wrong.

In my first company, I started with the assumption that content would be free and content companies would need a business model to stay in business. I thought advertising would be that business. We were just starting to see sponsorships on sites like Netscape and Yahoo!

Every week I would brainstorm with my soon-to-be co-founders about something, then go out and talk to every possible customer to see if they were interested. I tried 10 ideas over nine months, the last one being an adserver that would rotate ads on a site.

I knew I had it right when I walked in to Yahoo! And they asked me if I could install it tomorrow (at that point, I had no code). So my co-founders wrote it that night and we installed it the next day.

Our product market fit was strong. Lots of people wanted adservers. I was head salesman for the first few months and took advantage of that time to talk to lots of customers.

This is where the idea of a pivot really comes in. Most people associate pivot with the most extreme version, which is to completely change industries, solutions, etc. There are lots of other pivots around business model, customer focus, etc. that many founders don't make.

I realized in one meeting that what the customer really needed was a system that could rotate ads dynamically so that they could promise a certain number of impressions for each advertiser.

So we went off and wrote that over the weekend and moved our customers to that. That completely changed our business because now adservers became compute intensive so you had to buy several to match your traffic.

About a year later, the first SaaS adservers came out from people like DoubleClick. We built one in stealth and revealed it just after we went public, growing revenues from $6m at IPO to $40m 18 months later. We focused here on advertisers instead of website owners.

Amongst these changes to our model and customers, we made a ton of mistakes, but they were forgiven because customer need was insane. We were literally the first thing new Internet startups bought when they got funded.

In the course of 4 years, we probably tried about 100 business model and customer pivot-level ideas and 4 or 5 worked really well. That was before AWS, before SaaS, before open source. Things were sloggy. Now you can try 1000 pivot-level ideas in four years and you need to.

Next time, I'll talk about a modern method of need finding. It's now possible to what I did in four years, but in four months. I'm hoping you will try that in your startup.

I hope you liked this thread. I write about all of the issues facing early stage startups. Please follow me @jwdanner and please retweet this to help other founders.

And here is the product-market fit (PMF) framework I use with my founder's to intentionally experiment until they find PMF:…