We are extremely fortunate that millions of users have (rightly) placed their trust in Synchronoss’s Personal Cloud products to back up and protect petabytes of content. That is a privilege we do not take lightly, and it will remain core to what we do. The next phase of our product evolution augments that promise with unlocking the value of that content. Put differently, it is not just that we have successfully backed up 7,970 of my photos but, through the new lens, we store – and can resurface – wonderful memories I have of watching my kids grow up, fun family vacations, and laughs with friends.

Hopefully that notion resonates strongly with you, much as it did with the many users and potential users we recently surveyed. It’s also a notion which is somewhat “easy to say, hard to do.” As such, I thought I’d spend a bit of time explaining our approach. I’m an evangelist of the Build-Measure-Learn framework and I’m grateful that my team has been open to tackling this opportunity via that methodology. We’re gradually exploring how we can surface / resurface those memories. It might start, for example, with our Artificial Intelligence algorithm building a Best of the Weekend highlight for a segment of customers, which leverages date-/time-stamp to isolate photos from the weekend and Machine Learning models to identify quality of images, duplicates and near duplicates, things in the pictures, and other elements which we could use in our attempt to find users’ best photos. We’ll fulfill the “Measure” part of the framework by seeing what resonates and what doesn’t. Quantitatively, we look at outcomes – did we successfully surface memories that led to sharing that content with others, ordering prints, etc.? Qualitatively, we’ll use surveys and app reviews to assess what is and is not working. For the “Learn” part of the framework, I use a mash-up of Alice in Wonderland, Henry Kissinger, and Yogi Berra (who, fascinatingly, all riff on this idea) to avoid the trap of “if you don’t know where you’re going (a la, what you’re trying to accomplish), any road will do (a la, you can claim success for anything if there aren’t clear targets)”. Another nuance to the Build-Measure-Learn framework is that it strives to shorten the cycle time for that loop. In that spirit, we are not doing one monolithic delivery. Rather, we will bring a Minimum Viable Product (MVP) to market and enhance/tune it over time.

It’s crucial to point out that our Machine Learning is permission-based and working with anonymized data. Users’ trust is sacrosanct to Synchronoss and we would never do anything to compromise that. Frankly, it’s also how Machine Learning inherently works. It’s irrelevant to the engine that it’s Rob Weinstein’s photos. What matters, in my case, might be that there happen to be a lot of pictures of Halloween house decorations or a multitude of pictures at different baseball stadiums. That affords our app the opportunity to personalize the experience for both me and segments of users like me. From there, we can explore if deeper personalization resonates more strongly with our customers; might I find more joy in seeing photos of my kids’ first Halloweens or pictures of Halloweens where we have had piñatas (yes, that has occasionally been a part of our festivities :))? Analogizing back to the image at the top of this post, we recognize that fire can be dangerous, so we start with ensuring the safety and privacy of our users’ content. Whilst maintaining that security posture, though, we hope we can bring light into our customers’ lives by perhaps showcasing joyful times they’ve had around campfires with friends and family.