Wednesday, August 16, 2017 Investment

Last year, I made my first angel investment in a long time, in a company called GridX GmbH, headquartered in Aachen, Germany.

Germany is pretty unique in how much investment there has been in solar energy. Driving through some parts of Germany, you'll see solar installed on most rooftops. Combine that with offshore and onshore power from wind, and you'll see an unusual effect: Sometimes, there is so much renewable electricity being generated that the price of electricity falls below zero.

At the same time, a new enabling technology is entering the market: At-home energy storage. Tesla Powerwall is probably the best-known product out there, but batteries are actually a fairly commoditized business and there are many competing battery makers. Now you have power generation and power storage installed in homes, and now you have to connect dots: If energy prices are high and you're producing it or have it stored, you should be selling into the market. If energy prices are low or negative, you should be drawing power from the grid and storing it in the battery. This is essentially energy market arbitrage for homeowners.

GridX creates the connection between the battery owners and the energy market: Install their product called the "GridBox", connect it to your inverter, and watch the Euros roll in. GridX will sell electricity from your solar panels to the market at the optimal price point, and store electricity in your battery when there is economic opportunity.

There is, of course, the feel-good aspect of enabling carbon-neutral energy generation at scale. But the beautiful thing about GridX is their network effect: If enough GridBoxes are installed, all energy can be traded in their network, eliminating the need for any traditional nuclear and fossil fuel power plants.

If you live in Germany, you can order the GridBox today at

Monday, August 07, 2017

Now that the User Downloaded Your Shiny New App, You'll See Him/Her Churn Because You're Asking for Email and Password

One of the most frustrating things in building a new consumer app today for iOS and Android is the user churn you get in trying to get users to sign up or log in.

The App format is already circuitous: You have to get users to both download and then open your app. Just that open step loses 10-20% of users when they forget that they installed your app in the first place. But then you have to make them create an account with a username and password, where you lose another 40-60% of users that don't want to make up and remember a new secure password, or won't complete the email verifications. You could offer logins with Facebook or Google, which lowers churn to the 20-40% range, but then you'll have to get and keep refreshing OAuth tokens and may still have to make users create their own separate password later. Even for apps that mirror a web presence where users have already created a password, you'll lose 10-30% of users who can't be bothered trying to remember it.

When developers see these sharp dropoffs in their signup and login funnels, the obvious reaction is to want to build a "logged-out experience" which lets users see some of the content in the app, but then requires you to signup or login if they want to create or interact. This is a seductive idea, and you will get more logged-out users explore your app, but they are less likely to see value in the app because of the limited feature set, or graduate to fully engaged users they still have to climb over that wall. The danger of logged-out usage cannibalizing logged-in usage is real.

Why are logins still a problem? It's 2017! Well, security and privacy are hard. Apps are sandboxed and can't get access to a stable unique ID that could serve as a proxy for identity (the iOS IDFA and Android Advertiser ID comes close, but can be changed by the user). You can't just have the OS autofill email addresses and autogenerate passwords because then every app would want to capture your email address. (Although variant of this, in which Android and iOS create proxy email addresses and offer an OS-level password manager could be the closest we can come to an elegant solution.) Android password managers offer their own keyboard that can autofill credentials - but realistically, only nerds like me use password managers. Any simple-to-use solution has to be much more fine-grained: For example, for the case where you downloaded an app for a website you've already logged into, Apple offers associated domains that lets apps access Safari autofill credentials. But this is only for people who use iOS, browse the web in Safari (not Chrome), have entered their password for the associated site on mobile web or their desktop Mac, and use apps whose developer went through the trouble of setting up associated domains. That's a small set of users.

Signups and logins in mobile apps will continue to be painful for the foreseeable future.

Wednesday, June 28, 2017

More about Area 120

I'm excited to share more about what we've been up to in Area 120 at Google: Check out

Thursday, October 06, 2016

ETH Zurich Interview

My alma mater's alumni magazine did an interview with me about post-school experience of moving to Silicon Valley and starting some companies - you can read it here.

Monday, August 29, 2016

189 Mental Models

This is a great post by DuckDuckGo's founder Gabriel Weinberg: Mental Models I Find Repeatedly Useful. It's a list of 189 mental models that apply to various life and business situations, all in one handy-dandy list.