- Everpix 1.0 was completed mid-August, so right in the summer like initially planned
- iPhone and iPad apps finally approved on 8/30 by Apple after a 5 weeks delay
- An important psychological milestone was passed: we have our first paying customers
- A new provisional patent application was filed
The team fixed 1,800+ tickets in the past few months to finish Everpix 1.0. We don’t have everything we wanted of course, but it’s a solid “infrastructure + web + ipad + iphone + mac” offering we’re quite proud of. It has been available for the last couple weeks, and except for some data integrity issues in some user accounts which are being fixed, there has been no major issues reported so far.
The major hassle towards the end was getting the iPhone & iPad apps approved by the App Store review board. It took several weeks, and ultimately a meeting with the App Store VP, to get through it. The blocker was our subscription based business model, for which Apple wanted the usual 30% cut, but on top of that wasn’t allowing us to do automatically renewed subscriptions, a serious handicap for a cloud based offering. Even more of a handicap if you consider that a close competitor, Adobe Revel, is allowed to do it. The matter was eventually resolved, the Apple folks seemed to have liked Everpix, so we should be good going forward.
We now start to have more significant statistics in the new infrastructure ,which allows us to fine-tune our initial projections on variable costs:
- 2,600 users have created accounts
- Close to 8 million photos have been imported
- The average number of photos imported for users who imported at least 1 photo is 4,500 (and almost twice that for Mac users)
- As expected, the average photo size has been oscillating around 500 KB
- However, image analysis required more hardware than expected i.e. takes longer to compute
Final confirmation will have to wait a couple months from now, after we have shut down Everpix Beta and have our latest infrastructure bills, but for now this means that, assuming a 5,000 photo collection on average:
- We are still looking at $0.50 / user / month in infrastructure costs
- We are revising to $1.25 instead of $1 the one-time user cost for initial import of photos
Everpix 1.0 infrastructure is really different from the one for Everpix Beta: full resolution photos, brand new JPEG 2000 pipeline, new image analysis, dynamic views on users’ collections... and of course a subscription system. Transferring old user accounts from Beta to 1.0 would have required significant engineering effort and would have still required users to re-upload their photos. Instead we opted for a clean cut:
- Move Everpix Beta to beta.everpix.com
- Send an email to every beta user telling them about Everpix 1.0 and inviting them to re-create their account there through a simplified process
- Give a 25% subscription discount to beta users
- Shut down Everpix Beta on October 1st
Contrary to our alpha to beta transition, we didn’t want to pre-create accounts for beta users in Everpix 1.0. Now that Everpix is a paid system, we just didn’t expect the majority users to transfer and subscribe. We would rather avoid polluting the system with lots of unused accounts.
The results of the transition so far are:
- 10,754 beta users had at least 1 photo uploaded and were invited by email to upgrade
- The email open rate was in the 45%-66% range as reported by our email service - it’s quite possible the remaining emails went unseen or in spam folders
- 1,037 accounts were created by these beta users
This means a conversion rate from a free-beta system to a trial-then-paid 1.0 system (with a 25% discount) of 9.6% to 14.6% depending if you factor in the email open rate or not.
We’ve been running at least 2 weeks late because of the Apple delay in approving our apps. This means we’re now early September, we have TechCrunch Disrupt next week, Apple is about to obliterate the tech PR landscape with major announcements and the rest will be taken by other big companies trying to play catch up or defense. It’s therefore a bad time to try to launch a product.
The plan is now to let September pass, wrap up and integrate semantic tagging so we can be much closer to Everpix’s full vision, polish the product more and make sure the infrastructure is rock solid, and more importantly hire a dedicated and experienced person for user growth / PR / marketing (we have been actively looking for the past weeks). Then we really launch with Everpix 1.1 with a revised target of early October.
As you may remember, we filed 2 provisional patents last September: one on importing photos from emails in the cloud and one on our early image analysis and assisted curation. Both are expiring in September so actions were needed.
For the first one, we have found that few users were using this feature and also that some prior art existed, so we decided not to file an application and let it expire.
For the second one, after discussing with patent lawyer and key investors because of the associated significant expenses, we decided to file an application (US only), which should be completed early next week.
As part of Everpix 1.0, we also have brand new image analysis algorithms along with an innovative way of doing semantic tagging resulting from the last few months of research. To protect this critical IP, we filed a provisional patent last week.
We have early builds we started testing on Windows 7, but nothing we can distribute yet. Current target for a public beta version is in sync with Everpix 1.1 launch.