Recently OtherInbox announced Defender 2.0, the next evolution of the product we launched at TechCrunch 50 in September, 2008. The new version is much better than the original, but it is also very different. Getting there involved killing some sacred cows and making a company "pivot" from one focus to another.
It was a difficult decision and probably took longer that it should have. I hope that by telling this story our customers can better understand the choices that were made and other entrepreneurs can make better decisions for their own companies.
The Idea
It all started with a"big idea". The cure for email overload. We were going to make email work better. Empower consumers. Make marketing efficient. Expose spammers.
The concept was based on a technique that I had developed in college and that my co-founder Mike Subelsky was using as well. We were both geeks and the idea was geeky, but we were going to make it so easy that anyone could do it.
At the time we called the product OtherInbox but now we refer to the initial product as Defender. At signup you receive a subdomain based off your username - username.oib.com. You "defend" your email by giving each website a different email address at that subdomain. Amazon gets amazon@username.oib.com and Facebook gets facebook@username.oib.com. You can just make up email addresses on the the fly without setting them up beforehand. We provided our own web page for checking email that displayed a separate folder for each email address you created. It is faster to process your email when it is pre-sorted. If anyone sells your email address to spammers it is immediately obvious because it keeps going to the same folder and then can easily be blocked with a single click.
It was very clever. On the occasion that you block a sender it feels quite rewarding and powerful. Giving out an OtherInbox email address to a store clerk often produces interesting responses that make you feel "in the know".
We worked on it in stealth for about a year before launching it into beta at TechCrunch50 in 2008. Thousands of people signed up from the initial press, but the product wasn't viral and we were not spending any money on advertising so the growth was slow. We were positioned against Gmail as the alternative solution and many people would respond that, "I don't have a spam problem" or "I can't believe it will be better than Gmail".
Most of the users who took the time to sign up and give it a try were very happy. They felt empowered and enlightened. They raved about it to their friends, wrote blog posts, and talked about it on Twitter. There just weren't enough of them! We weren't seeing the kind of adoption that we had hoped for. Certainly not enough to make our business model work.
As Josh Kopelman pointed out to me at the time, there were a lot of "speed bumps" in the way of becoming a Defender user. You had to:
- start over with a new email address
- learn a new way to give out your email address
- check your email in a new place
- invest hours of time in switching to the new system
- wait weeks before really seeing the benefit of your effort
- targeted at power users
As much as it was cool to block senders who spammed you, it didn't really happen very often for most users. In fact, very few users had ever clicked the block button. The idea of having a folder for each sender worked well with dozens of senders but broke down over time as you accumulated hundreds or even thousands of senders.
We racked our brains about how to make it easier.
We came up with a sophisticated "import wizard" to get you started. It would connect to your existing email account and suck in all of the email that you SHOULD have given out an OtherInbox Defender address for (the stuff that's not from real people) and organize it into folders. Then we were going to write a screen-scraping system to change your password for you automatically at all of the different sites. Besides being a maintenance nightmare, this would have required you entering your password for every website you have a subscription at into one of our forms - something that I'm sure not many people would feel comfortable doing.
We build the first part of the import wizard that sucked in your email messages from Gmail and right about that time we started talking to Yahoo! about how we could adapt Defender for Yahoo! Mail. Hence our second product, Organizer, was born.
The Pivot
Where Defender worked off the address it was delivered TO, Organizer worked off the address it was sent FROM - that way users don't have to do anything different. Yahoo! doesn't support nested folders so the idea of creating a folder for each sender was impractical and we wanted to try and keep the user experience inside of Yahoo! Mail as much as possible. To simplify things, we decided to categorize each sender and only create a folder for each category. This way each user would have 5-10 OIB folders instead of 50-100 or more.
The Organizer product really took off at Yahoo! Mail. In less than a year we signed up more than 500,000 Yahoo users with very high retention rates. Something was working! We created a Gmail version of the Organizer and it fit in great with the IMAP/OAuth/OpenID model that evolved with the Google Apps Marketplace.
The positioning of Organizer resonates with a wide range of users. It's not for spam, it's for emails you asked for such as Amazon receipts and Facebook notifications. It doesn't replace your existing email account, it makes it even better.
Everywhere Defender had a problem, Organizer addressed it:
- you keep using the same email address
- you don't change the way you give out email addresses
- you keep checking email the same way
- it only takes 60 seconds to sign up
- there is no training or setup required
- your instant gratification is that your inbox immediately gets smaller
- targeted at the everyday email user
At first, we thought of Organizer as an inferior version of Defender. Soon we realized that Organizer was actually better than Defender for for most people. We found that most people don't have spam problem - the spam filter at Gmail and Yahoo is good enough. Most people do have an email overload problem - they get too much email they actually check the box and ask for. Most people do not want to invest time in improving their email or learning to do something different. Organizer helps anyone who gets too much email without them having to do anything different - the magic combination!
We had to pivot and focus our attention on Organizer, even though Defender was our first love.
It took too long
Nobody wanted to hear this. I certainly didn't want to admit that the "great idea" we started with was not going to be successful. We probably should have began this process six months earlier than we did but it was hard to swallow.
That put us in a situation where we had more than 500,000 users signed up for Organizer and less than 20,000 for Defender. Only about 1000 were paying $20/year. Yet Defender was about half of our code base and 80% of our customer support issues. It was hard to put many resources into improving Defender because 95% of our users were using Organizer. Yet everyone on the team believes in Defender and uses it ourselves. It hurt the morale of the entire team to see a product stagnate from lack of attention. We needed a way to combine our resources so that all of the effort we put into the product can be benefit both Defender and Organizer users and so that we could deliver a more reliable service.
There was a lot of duplicate code:
- web UI
- imap
- mail receiving
- rss
- message display
None of these items are competitive advantages for us. We're never going to make a better, more reliable IMAP server than Gmail and it wouldn't add much value if we did. If we could find a way to use Gmail or other email services for these common tasks and add our value on the Organizer and other unique services it would be a big win for everyone.
We debated it and debated it. Emotions were high. Making big changes to Defender felt like quitting or admitting defeat. We didn't want to let down our first users, the early adopters who believed in us from the beginning. We hated seeing Defender stagnate from lack of resources and wanted to give them the new features we had developed in Organizer.
Which leads us to Defender 2.0. We came up with a way that I believe gets the best of both worlds. You can still give out multiple email addresses and block certain senders if needed. But you also get the benefits of great search and other features from Gmail. Defender 2.0 is designed to work with Organizer, so that users also receive all of features of that product as well.
Defender 2.0 on Gmail brings:
- search
- a single OIB label to view all mail at once
- custom filters
- send message attachments
- view all message attachments
- modify forwarded messages
- more reliable IMAP
- more reliable RSS
- and much more
I've been using it myself for six months and would never go back. I still give out a different email address to every website that I register with. I love being able to use Gmail search and to have OtherInbox as an additional email account on my iPhone. It is faster and more reliable than before.
There were some sacrifices though. I really like the sender view of the original Defender and Gmails filtering options aren't as robust as I would like. If I want to send email from amazon@yourname.oib.com extra work is required to verify that email address.
Some of our most loyal users got really pissed off! They felt let down. Both free users and paid users alike complained that they were going to miss aspects of Defender 1.0 or that copying their emails over IMAP was tedious and hard. This was very hard on our team as well. We wanted every user to see this is as an improvement. We wanted everyone to be happy. We will refund the payments of paid users who aren't 100% satisfied with Defender 2.0 but getting a refund isn't going to make anyone feel better about it.
Lessons Learned
It's hard to admit that something isn't working. It usually takes too long. But when you finally do it, a great weight is lifted.
When have you admitted that it wasn't working? Please share your examples in the comments.
Recent Comments