Over the last few months we've been working on landing a licensing deal for Zoto in the professional and college sports arenas.

On June 1st, Zoto entered into an agreement with the Charlotte Bobcats to provide a photo management system that will allow them to upload, store, organize, publish and print the photos in their repository. The project is to be deployed by the time the NBA pre-season starts at the beginning of October, and will provide facilities for taking photos from fans, and republishing the photos into media-rich albums hosted on the NBA servers. Zoto will be relying on printing from it's partner Qoop, who provides quality, user-customizable printed products.

It's an exciting prospect for us here at Zoto, and marks significant shift away from our consumer based business model.

While we have enjoyed moderate success in the consumer sector, there are simply too many companies competing in this space for us to make a noticeable dent in market share. By going vertical in the photo market we have been able to find a niche where there are very few competitors, a definite need for a better product, and a large demand by users for the images we will be storing. In short, the licensing model makes better sense for us than continuing to compete in the consumer market.

Zoto will continue to run the Zoto.com site for consumers, and the great news is that Zoto.com users will continue to get advanced features for managing their photos as we develop them for our licensing customers. Many of the features needed by this new model relate to features already desired by the individual consumers, and we will be integrating these features into Zoto.com where appropriate, and once they become available.

We would like to thank all of our users (free and paid) for using our software over the last few years! As Zoto.com remains our showcase for most of the features we provide, be assured that we will continue to prioritize feature requests and bug reports by the users.

Today we'll be pushing out the last of the code that completes the optimization work we started over a month ago. This last portion of code is a substantial rewrite of the JavaScript that Zoto uses to display pages on the site, and will finish up our work on speeding things up.

Before today, each time you went to a new page on the site your browser would have to re-read the JavaScript we use on the site from your browser's cache. Because of the size of the code we use, some user's browsers were slowed down by this reading in and parsing of the JS code. The changes we have made to the site now allow your browser to redraw the different pages without having to reload the JS each time, which effectively cuts down page draw times. The net effect of these changes are that pages now 'pop' in almost immediately, with no discernible delay in moving from page to page.

There are some caveats to this approach. First, the forum, blog, albums and image editing pages will not participate in the speed boosts. If you navigate to one of these pages and then back to the site, there will still be a small delay as your browser reloads the JS it needs to draw the page itself. However, these pages are used by the users infrequently at best, and should not appear to be making things 'slower' than they already were before today.

Even though we have tested this extensively on both development and staging, there will be some bugs that appear as a result of these code changes. As always, we remain committed to addressing these bugs in a timely fashion, and expect to have any issues resolved by early next week at the latest.

Our work now shifts to focusing on album features, including ways to secure your albums, creating new templates for use on albums, and beginning work on the community sections of the site.

Please post in the forums if you see anything out of the ordinary with these changes, or want to comment on any speed boosts you may see!

The site has been getting steadily busier since our release in March. This means we are doing more work serving pages, and certain portions of the site have become slower as a result. However, over the last 2 weeks, Josh has been working on optimizing our backend infrastructure to speed up things significantly.

Pages and page elements (like the tag cloud) on Zoto were taking a considerable amount of time to load due to the number of queries being done on the database server here at Zoto. The net effect was a slowdown for page load times on the site, which exasperated the problem with JavaScript load times (see post below).

To fix this problem, we have implemented a couple of new technologies for speeding up queries on the system.

The first technology is memcached, which was written by the guys at Danga Interactive for LiveJournal. We are now using memcached for caching things like user information (username, password, email address, number of messages, etc.) and storing those results in server memory for rapid access. Now when you move from page to page on Zoto, your account info is pulled from server memory, dramatically decreasing the amount of time needed to display the new page for you.

The second piece of technology we've implemented is a technique for databases called materialized views. Materialized views are 'snapshots' of data in our database that may get used on a regular basis. We index these views so that when you access particular portions of your account, like the lighbox, we can deliver things like the tag cloud to you almost instantaneously.

Both of these updates have had a positive, noticeable effect on the server loads here, and have significantly speed up the rate at which we can serve content.

We are still left with the problem of how to speed up JavaScript rendering times on the user's browser, but we have a couple of ideas on how to fix this as well. Within a few weeks we should have the majority of the speedup features implemented, and any noticeable delay (even on slower computers) in our page load times should be gone for good.

Over the next few weeks we will begin working on optimizing the Zoto site to make it "faster" for the users complaining of it being "slow". However, I must initially point out that Zoto's site itself is not necessarily slow - the slowdown comes from how we have implemented our software, and how fast that software runs on particular computers.

To understand this seemingly conflicting information, one needs to understand how Zoto's site actually works. When a user hits the site, Zoto downloads a bunch of code to the user's browser that tells it how to build the page they are looking at. Very little data has to be sent from our server to the user's browser to make this happen - and in most cases, the user's browser will have a cached copy of the code that does this. The pages themselves are built from a) a bunch of JavaScript (JS) code we wrote, and b) a bunch of data that comes in from the server, which is in turn pulled out of our database.

In simplistic terms, if Zoto seems slow to you, it's because either your browser is slow at rendering pages with JS, or your computer is slow and takes longer to build the pages than a faster computer.

For example, on my computer at work (a year old Apple G5) our pages display faster in Webkit (an optimized version of Safari) than Firefox, and display faster in Firefox than a plain Jane Safari. My laptop (a 2 year old Apple G4) displays pages on Zoto slower, because it's slower than my desktop.

To understand this better, one needs to realize that users will perceive a page is slow based on how fast information initially appears on the page after clicking on a link. A user will say a page is "slow" if that page fails to display anything within a few seconds of clicking on a link. If anything appears on the page within a few seconds, the user will usually tolerate the rest of the page being loaded.

Our problem with this is that our entire page is built with code running on the host computer. When that computer is slow, the pages don't display immediately - instead the host computer starts running the code first, then starts building the page after that. In addition, different browsers will display portions of the page at different times - so IE6 may wait till the entire page is done building before displaying ANYTHING on the page, where Firefox will display bits and pieces as they fill in.

The way to "fix" this problem is for us to start giving user's browsers more information for building the page up front. If we do a portion of the work needed to render a page on our server first, the host computer can "fill in" where it needs to, but the net effect will be the page itself will appear to load very quickly. As we have only a few primary pages that need this sort of love (user homepage, lightbox, image detail) it should be straightforward to implement.

Additionally, we are implementing a technology called "memcached" on our servers. We already implemented portions of this last Friday in a few simple areas, but we will be adding it to most of the calls our pages make back to the servers. This will help eliminate any periodic slowdowns attributed to making calls for data (like the list of photos in your lightbox, or the widgets on your homepage), which appear to make the pages slow to load.

Optimization is always an ongoing effort. Sites that have been up for years go back and optimize their site when new technologies make things faster for the users. Zoto is no exception to this rule, and we'll continue to strive to improve the performance AND features of Zoto as we grow.

I'm attending the Web 2.0 Expo in SF this week, and I'll be posting when I find something worthy of discussion.

The first talk is done by Steve Souders and Tenni Theurer from Yahoo. Both Steve and Tenni work at Yahoo's Exceptional Performance team and they have identified several factors that affect the load times on Yahoo's properties.

Testing Your Users
One of the tests they ran was a load of a 1x1 GIF to test whether or not people have the cached copy of Yahoo's content on their computer - or not. It was a shock to them that they saw 40-60% of the visitors were visiting with an empty cache.

I asked whether or not they measured the breakdown of browsers on this test, and they had not. I think it might be beneficial for Zoto to run this test - with browser breakdown enabled. I suspect that Yahoo's results might be flawed due to the lack of a breakdown of browser type. IE6, for example, has many bugs related to its cache design and may be influencing the results heavily based on its share of the browser market (about 35-40% of Zoto's traffic).

Multiple Browser Connections
Regarding using multiple host names (as we do) for speeding up response times for loading images, Yahoo found that using more than 2 hosts (like www1 and www2.zoto.com) caused slowdowns due to thrashing of the user's CPU trying to load more images.

This could cause browsers such as IE6 a considerable amount of issues when loading more than a few images at one time.

We are currently using 4 names (www1, www2, www3, www4) for serving thumbnail images, as well as using 'www' for larger images. The use of 'www' for the image detail modal images allows that image to load before the thumbnails are finished loading if the user clicks on a thumbnail. We may want to try using 2 names (www1, www2) to test if this helps out the IE6 users a bit.

Combining JS, Images, and Style Sheets
One thing we've learned is that combining JS files is important in cutting down the time it takes to connect and load each separate file loaded from the server. Yahoo's suggestion (although they aren't using it themselves much of anywhere) is to include your image files inline in your stylesheets. By Base64 encoding the images for your site, you can stick them in the CSS file, which is a single download.

Since we've pushed the color switcher, this is now a good possibility for us to speed up our page loads. The caveat is that (go figure) IE6 doesn't play nicely with this technique.

It's now been about a month since we launched 3.0. Things have calmed down a little bit since the launch, and with fewer and fewer issues being reported we are now moving out of fire fighting/bug squishing mode. I figured it would be a good opportunity to tell about what we've been doing regarding moving forward with Zoto and our new paid only model.

Overall, I'd say most of the existing users seem quite pleased with the new system. Of the few complaints we've seen, most of them have to do with something we took away, like the album templates, or changes in the way things look, like the new pink based theme. New users seem to be the most satisfied, as they have no preconceptions based on the old site. We remain focused on delivering desired features to ALL users regardless.

One interesting side effect of going to a paid only model (other than we just got rid of 99% of our old 'users') is that the number of photos in the system is growing at a rate just under or equal to the uploads seen before the launch. This comes as no small surprise as 22% of the photos we pulled into the new system on launch day were owned by less than 1% of the accounts on the system. With the new uploader rocking and rolling along, and the speed and ease at which people can now access their photos, it makes sense that we are seeing the number of uploads we are.

Unique visits to the site are down slightly, but not as dramatically as we originally projected. We long suspected that most of our traffic came from users using us to host static image content for free from other services that charged for bandwidth. Now that we have eliminated those freeloading accounts, we see our bandwidth usage a bare fraction of what it was before, yet the number of simultaneous users using the system seems to be about the same - irregardless of the fact we got rid of 99% of the older accounts.

Obviously a side effect of the change in our model has been the number of raw signups (free and paid) we are seeing. Again, because of the fact we got rid of lots of users, and that those users made up a particular demographic (people wanting 'image hosting'), this is to be expected. Ongoing, we plan on offsetting the decreased number of signups with paid ads on Google.

When we ran ads previously on Google, we saw poor conversion numbers, with advertising dollars being spent that could never be recouped with actual paid revenue. Roughly speaking it cost us about twice as much to get a paid user as we got from them signing up for a paid account. While we saw a HUGE number of free signups with those ads, those accounts were worth little to nothing (if not less than nothing) ongoing.

Time will tell how our new ads perform as we are just now starting to tweak the offerings we have on the site and the wording of the ads we are running. However, we fully expect to be able to directly add a paid account to the system for less than it costs to acquire them. Our cost of providing service to a user (CoGS if you will) moving forward is fairly minimal, so throwing money at the problem of signing up users should be a straightforward as spending ad dollars with Google, and making money off the new signups.

We have recently discussed licensing our software to several fairly large companies. It may turn out that this revenue model provides the real meat of our revenue, with the added benefit of giving our existing website property zoto.com features that come about from these licensing deals (new templates, features, color schemes, etc.).

A few pessimists have seemed quick to label our moves as a 'bad thing', but we feel that with such a substantial improvement in our offerings, coupled with our new subscription model, we are finally on the right track to success.

Kara pointed me over to a blog post about Zooomr today. She half kiddingly said that I should go off and author a post about why Google should buy us instead of Zooomr. I skimmed the article and shrugged it off as either a) someone who is clueless about why Google really buys companies, or b) a shameless 'self' plug. And, frankly, I don't want anyone to buy Zoto nowadays. I want us to make it on our own merits, and not have to rely on someone 'rescuing' us by acquisition.

I've been aware of Zooomr for some time, but it was only tonight that I decided to head on over to their site and check out what could be so worthy of being acquired. Imagine my surprise in finding screenshots that look suspiciously like Zoto 3.0 on their blog. WTF?

Now I'm usually a good sport about this type stuff. I've posted before about how Flickr has a habit of picking up features of other services and integrating them into their offerings. For example, I would suspect that Flickr's new homepage layouts, and set collections have a bit to due with our new lightbox designs and album sets. It only makes sense they saw our new stuff and said "hey, we could do this with ours!". Regardless, while this is probably an ego-centric view due to the fact that I work at Zoto, it's good to see the guys at Flickr are continuing to think about how to make a better product and how to make that feature "their own". We all do it to some extent, but the key factor is that we don't rip each other off blatantly.

Perhaps not so with the Zooomr guys. Here's a screenshot of what I saw on their blog:



The image on the right is showing what we call an "image detail modal", but on Zooomr it is simply them using Lightbox JS v2.0. On Zoto that lightbox modal is 100% our code using Mochikit for effects - nothing else. The shot on the left is what I really have an issue with - it looks almost exactly like our own new lightbox in minimal mode. Not only is there a view mode switcher at the top to switch sizes, but the buttons below are almost exactly like the ones we use on our slideouts in bulk mode. For me the dead giveaway was the 'heart' button, which we had on the Preview of Zoto 3.0 for about 5 months, before we took it out.

Here's a early mockup out of our dev-wiki from mid-2006 that shows a variation of the heart/favorites button:



Now maybe I'm wrong about all this, and maybe Zooomr came up with it themselves independently. There are things on every photo sharing site that are needed and common to each, like a prev/next image, or a page with a bunch of thumbnails on it. However, I will point out that we've had a working lightbox like what is on 3.0 today for about 12 months now. It was almost a year ago that I showed a rough version of this off at a conference in Mountain View, and a working version has/had been up live on www.zoto.net since Sept/Oct of 2006.

We've spent a lot of time and money working on this product. While innovation by imitation is certainly fair game, a knockoff is simply inconsiderate and shows a lazy approach to development. If you can't design your own interfaces and write your own code, you have no business trying to DO business in this space!

Zooomr seems to get it's fair share of press regarding their innovations. We here at Zoto are a little more laid back about publicity, choosing to worry ourselves with writing better code. FWIW, we have an in-house designer, Kara, that did all the interfaces and designs, and we wrote 100% of the backend code for hosting the site, and 100% of the code on the front end, less the Mochikit stuff we use to do event handling and effects.

It is my hope that other companies that consider copying other's work also consider the ramifications of doing so. While it's great that certain young people are capable of developing a lot in a short amount of time, it does NOT make it OK for them to take shortcuts at another company's expense.

With age comes wisdom, and the realization that consistent hard work and effort are where the real rewards lie.

Whew. That was a crazy 5 days of scrambling around trying to get things running normally on the new site.

We've exported 99% of all the pro accounts to the new site. We've also got all the old free users in the system, and have it set up to allow them to upgrade to a paid account for $9.95.

If you are one of the few users we didn't manage to import into the system due to data issues, we apologize. We will be working this week on importing you and squishing bugs we and others find.

I'll keep this short as we're all beat, and need to sleep, as evident from Clint's hair.

We are still in the process of working on the Launch of Zoto 3.0. We estimate that the new site should be up and operational sometime in the next 24 hours.

Exporting and Importing
We are currently exporting and importing paid accounts from the old 2.0 system into the new 3.0. As things would have it, the process of importing users into the new system is taking a bit longer than we projected initially.

The good news is that most of the paid accounts exports are finished, and should be live by the end of the day on Friday. The bad news is that a few pro accounts may have their import delayed until we can finish processing the backlog of imports.

Prioritizing Account Imports
We are prioritizing accounts based on the number of photos in the account, and the date of the last login. Essentially, if your account has a smaller amount of photos than another account, or you have logged more recently than another account, your photos will be imported in the next 24 hours.

Additionally, we are putting into place a priority for handling the users that will log into the system when it goes live.

Basically, if you login to Zoto in the next 48-72 hours, and you don't see your photos in your account, the system will put a priority on importing your account data for you. You should see your photos within a few hours of your first login to 3.0.

As soon as we have a better idea of the launch time, we'll post back here.

Hang tight, it's going to be well worth it!

Pete Cashmore over at Mashable seems to think he's the all seeing eye for Internet company's business models.

Pete, you simply have no clue what you are talking about. Making the statement that getting rid of free users is a sure way to kill your business just shows your naivety when it comes to sound business models. Let's hope you aren't in charge of anything financially important wherever you are.

With a free/paid model, at its best, we were converting 1.5% of our users to paid accounts. To help drive that conversion, we had to restrict quite a few features for free accounts, and constantly tweak our ads that we were running on Google.

Now I know that 1.5% conversion isn't great - it was due in large part to the offerings we had on Zoto 2.0 - most of which weren't that impressive when compared to other services. It goes without saying that revenue wasn't impressive either - assuming that we stuck to the free/paid model.

I've had the opportunity to talk to people involved in other companies similar to Zoto, and it turns out that their conversion rates weren't all that high either, with the largest of them being about a 5% conversion to paid accounts, and a good average being around 3%. Just imagine you were a car dealership, and 97% of the cars you 'sold' were given away for free. It would totally suck.

The consensus on why these conversions are so low for free/paid models is that if you offer the option of a free account to people visiting your site, you're going to have a LOWER paid conversion rate simply because people will take something for free over paying for it.

At the end of December we turned off the ability to sign up for a free account on Zoto - within hours of doing so we saw our paid signup rates jump to 5 times what they were before turning off free signups. The reason why it jumped is because, given no other choice, some people will elect to pay for a service. That rate has remained consistent to today, even with the old tired site still in place.

As the comments on Pete's post mention, SmugMug runs a paid only service and has been quite successful with their business model. I'd estimate that they have around 200K paid accounts, and at an average of $40/year each, they should be making around $8 million a year.

On the flipside, Flickr does an admirable job at converting users to paid accounts - for one they restrict the amount of photos uploaded per month for free accounts, and the number of sets you can create. Based on that, let's assume for a moment that Flickr has a liberal 7% conversion rate.

With a guess of 5 million users to date, Flickr should have about 350,000 paid users. 350K x $24.95/year = ~$8.7 million in paid subscriber revenue per year. I won't bother with guessing what ad revenue is - it's going to be a smaller fraction of their income regardless.

Flickr's pages show they have 29 people on their 'team' at Yahoo. Assuming they pay each of these team members an average of $100K a year, that's $2.9 million in salaries. That leaves Yahoo with a 'guestimate' of $5.8 million a year in revenue, before CODB, on the Flickr property.

Flickr is one of the most popular websites on the Internet. I'd be willing to bet anything that their cost of doing business (hardware, bandwidth) is well in excess of $10 million a year, simply because they are providing services to all those free accounts, and those free accounts suck hard drive space and bandwidth like nobody's business. Remember, that's 93% of their users they have to support for free.

FWIW, I think it's great that Flickr got acquired by Yahoo - it helps ensure that they'll be around a few years from now. Yahoo has enough revenue from other sources to keep them going.

For us however, and the likes of SmugMug, our future lies in creating a real, viable business model - one that makes the company money, and provides enough revenue to ensure the PAID users get a quality service, without having to scramble around and worry about getting acquired tomorrow.

Simply put, basing your company's model on 'cramming as many free users into your company as fast as you can so you can get acquired' is a supremely stupid business decision.

The day and age of doing Internet startups in the slim hope of becoming a 'YouTube' is coming to an end. May the rest of us with our sanity intact enjoy working at a healthy, profitable company.

It's time to launch Zoto 3.0!

Signups, logins, and upgrades on Zoto have been suspended in preparation for the migration of the user data to the new system. The Preview of 3.0 is also being shut down today in preparation for the launch. The final step of shutting off logins was done at about 3PM CST.

Account Transfers
All existing paid (pro) accounts will have their data imported into 3.0 today and tomorrow. When we bring the new site up tomorrow, paid accounts will have complete access to all of the older account data, and should be able to start using the new site in it's entirety.

As of launch time tomorrow, all free accounts will be required to upgrade their accounts to a paid account, if they want to continue using their account. The cost of this upgrade will be measly $9.95 for a year's service. That's a great deal, and will be offered for a limited time only.

If you elect to NOT upgrade your account, you will have the ability to download your photos and meta data (title, desc, tags, etc.) on or after the 2nd week in March. This delay is due to the time frame required to export and archive your account data on the system, and has absolutely zip to do with wanting you to upgrade!

Note: If you need your account data sooner than the 2nd week in March, you can either a) upgrade to a pro account, or b) send us email at website-support@zoto.com requesting an archive to be posted for your account. Depending on the request load, we will process this as promptly as we can.

Welcome to Zoto's new blog! This blog post is also mirrored over on our old blog.

The Zoto.net domain is currently running the Preview of Zoto 3.0, which is due to launch in March of 2007. If you are interested in checking out the new features of Zoto 3.0, you can signup for the beta test for FREE right now. If you just want to look around at some existing accounts, you can visit the community pages. You can also check out activity on the new site by going to our community spy page, or discuss the new site on our newly revamped forums.

Pre Launch Christmas Special
If you sign up or convert to a paid account between now and March, 2007, you will get an extra year FREE when we launch 3.0 in March, 2007. Your account will automatically be upgraded when we launch the new service and you will not need to do anything extra when we upgrade.

Changes to Existing Accounts
Starting today, December 22, 2006, Zoto is moving to a paid-only revenue model. Free accounts are no longer available to new users of Zoto. We made the decision to move to a model that provides us with a way to continue operating Zoto without relying on an exit through acquisition. We feel that decision is the best way to move forward with Zoto's growth plans.

Here's how the upgrade to 3.0 will affect existing and new accounts:

• If you signed up for a "Pro" paid account on Zoto prior to 12/22/06, you will automatically have your account upgraded to a 3 year account for no additional cost when we launch 3.0 in March. Your account will be moved over when we upgrade, and you will need to take no further action.


• If you signed up for a free account on Zoto prior to 12/22/06, you will be able to upgrade your account for a minimal cost when we launch 3.0 in March. Your account will then be scheduled to be moved from the older servers to the new release. You may continue to use Zoto 2.0 (on Zoto.com) for free until we upgrade to 3.0.


• If you signed up or converted to a "Pro" paid account after 12/22/06, you will automatically have your account upgraded to a 2 year account for no additional cost when we launch 3.0 in March. Your account will be moved over when we upgrade.


• If you sign up for a paid account on Zoto after 3/1/07, you will pay $29.95 for a 1 year account.

Exciting New Features
From the beginning we designed Zoto to be a place that you took ALL your photos, not just the few you wanted to share with the world. Now, just like a real photo sharing application (think iPhoto or Picassa), the new Zoto excels at providing you a lightning fast way to organize, view and share your photos.

What we've ended up creating with 3.0 is a true web-based application for managing TONS of photos. Because it's web-based, you'll have access to all you photos from anywhere in the world - at work, home, or on vacation.

We're not going to try to cover all the new features we'll have at launch, but here's a short list to get you fired up:

• Customized user homepages with drag and drop elements, and the ability to create groups of photos, feature favorite images, and display friend's photos, comments on your photos, and tags.


• Easy to use tabs and menus, and a powerful search feature that allows you to find photos in a snap.


• An advanced lightbox that can show up to 160 photos on the screen at once, with inline modals to show you larger size images without having to navigate to another page to view them.


• Slews of different views for the lightbox and the ability to sort images by date, camera make, f-stop, iso speed and more.


• An integrated "bulk mode" in the new lightbox, allowing you to edit, tag, publish, export or print 100s of photos all at once.


• A new "image detail" page that shows HUGE images, and provides powerful editing of photos, including cropping, saturation levels and more.


• Advanced permissions that integrate with your contact lists, allowing you to control EXACTLY who sees what image.


• Tons of other features too numerous to list, including new customizable galleries, flash export widgets, and more.

We are really, really, REALLY excited about our new release. It's been a long time in the works, but we're positive the wait will have been well worth it. If you have any comments or questions, you can contact us by email or by posting in the forums!

Have a great holiday season!