How to Get Paul Graham to Beta Test Your Startup Idea

aka How to Defy Google Using Eric Ries’ Lean Startup
by Eugene Leychenko and Greg Brooks, Founders of


We hated how Google was making Gmail less user friendly - Motive
We listened in on what people disliked about changes to Gmail - Customer Development
We created a lightweight product as quickly as possible - Product Development
We created a viral loop - Marketing (with a bit of friction)
We went out of our way so our users could speak directly to us - Ongoing Customer Development
Re-released with all the fixes and now are ready for primetime. - Giggity

Around mid-August 2013, Google made drastic design changes to Gmail. The most controversial change was shrinking the size of the compose window to moreso resemble a chat box rather than an area to write emails.


We hated it. Our friends hated it. People even made petitions against it. We wanted to do something about it. We figured there must be some way that we could get our Old Compose back.

After a weekend of research, we found a solution that would return Gmail to its former self! However, we wanted to move quickly. Very quickly. Like go-from-inception-to-full-launch-in-a-week quickly. Of course, we had no idea what the expectations would be. We just knew we needed to get going.

So, we began coding a Firefox and Chrome extension that would let users revert back to their Old Compose. Simultaneously, we started advertising on Facebook to target demographics in order to assess the general reception for a productivity tool that would take Gmail back to its Old Compose settings.

Originally, we planned on charging a nominal fee of $10 or so.

That plan changed pretty quickly when the Facebook ads results came back. Users interacted with our ads at a surprisingly high rate. However, they also sauteed us for seeking to charge for our service. It seemed that people were not comfortable paying for a feature based off a free product. While we couldn’t be sure this was the majority view, we were focused on speed, so we made the quick decision to launch our beta for free. This was the third day of our adventure.

Within twelve hours of determining to launch for free, the physical product was completed.

Next, we mocked up a website (in less than a day) using LeadPages. Instead of asking for money, we decided to allow users to get the extension for free by Paying with a Tweet. We hosted the extension on our site as opposed to the Chrome webstore/Firefox add-on store, and we also asked for the email addresses of our users. Our thought process was that we didn’t know what direction we would eventually move towards, but we wanted to retain access to our users.

These events marked the end of day 5.

We spent the first half of day 6 making sure that everything functioned properly. Then, we sought the final piece of the puzzle…

PR, ideally from a major blog.

We quickly realized that Lifehacker would be the perfect avenue because of it’s early adopter, tech-oriented audience. On the evening of day 6, we fired out a few emails and offered Lifehacker an exclusive.

We woke up on day 7, September 13th, 2013, to a front page Lifehacker article.

The results were interesting. Predictably, we sat by our counters all day and watched our user numbers rise (very exciting!). However, we learned that we created a floor by requiring everyone to tweet about us. Inevitably, someone would see a tweet, and then sign up, and then tweet, and so on. We verified these suspicions through GA, as we were getting lots of Twitter traffic.

Here is what Twitter looked like in terms of people mentioning “@OldCompose”, almost exclusively due to the Pay with a Tweet campaign:


People really loved the extension because it did what we said it would do - it brought back the Old Compose.

A major factor of why we built the product was because people were so vocal about their dislike for the changes to Gmail. This tweet by Paul Graham stood out:

Seemingly out of nowhere, one of our early adopters (Thanks Jason!) wound up telling Paul about the product!

He even tried it out, and, to our dismay, did not respond positively: After assumably drowning our sorrows in ice cream, we concluded that Paul didn’t like the slight friction that free advertising creates. However, we felt like we were already playing with house money since we were only a couple of weeks into the project and had already received more press and discussion than we anticipated.

That was the product of applying Lean principles. We decided that we would take a mere seven days to give the extension a go. We asserted that the petitions, vocal social media posts, and various Google results for “I hate the new compose” indicated enough demand to see what would happen if we created a product (note that this is a heck of a lot different than if we were launching a full scale business).

We also divided and conquered — one of us headed product development while the other tested the Facebook advertising waters. This strategy turned out to be crucial; if we had, for example, launched with a price tag, it’s very likely no one would have written about us.

We stayed the Lean course by laughing at the idea of building a website, instead using a glorified text editor and the ever-useful inspect element to do the heavy lifting.

Lastly, the whole time crunch in general was critical to a successful launch. By agreeing only to invest a week of our lives, we were forced to move quickly and make decisions off incomplete data; but it also lowered our expectations. While we maintained full effort as per usual, it was “okay” if everything didn’t go swimmingly. This mindset made it easier to shrug off negative tweets from people with 200,000 followers, for example!

Over the next few months, we created a culture where our users had an open door to us: they could tell us what they liked or disliked about the product, what bugs or incompatibilities existed, and how they wanted to mold future iterations. A key theme here was inconveniencing ourselves in exchange for a convenient user experience. We answered a lot of emails with similar questions, or incomplete questions, or where a knowledge base would have provided a solution. However, we wanted to lower the barrier for users to communicate with us as much as possible.

Now, while incorporating a good amount of that user feedback, we’re ready to re-launch with Old Compose 2.0. Interestingly enough, Google wound up breaking a bunch of our competitors via routine tweaks and adjustments to Gmail. In response, we re-built Old Compose from scratch so it would be immune to Gmail changes (and also so it would function with all major extensions).

We’re now ready to take Old Compose out of beta and release it to the world. Thanks to our Lean approach, we’re pretty damn confident you’re about to become a happy user!