The Pinterest Roadmap

The rise of Pinterest wasnt all cocktails and cupcakes. It conquered huge scaling challenges to become the worlds keepsake box. Now as the we ditch the desktop for mobile, how will Pinterest evolve to connect us with the things we love? Today Pinterests head of engineering Jon Jenkins gave a surprisingly candid look at the past and future of Pinterest journey to build the interest graph.

Now Pinterest is a team of 140, with 70 engineers, $338 million in funding, and a massive headquarters in San Francisco. But at the start, it was just the three co-founders Ben Silbermann, Evan Sharp, Paul Sciarra, and a single engineer.

Pinterest began as a few guys with a mission To help people discover the things they love and then do those things in real life. With only one engineer,Our top picks for the cableties and gear, it was almost lucky that Pinterest didnt become popular yet or the whole thing could have come crashing down.

Jenkins says that traffic was doubling every month and half and provided this traffic graph, though with no Y axis about exactly how many page views the site receives.Weymouth is collecting gently used, dry cleaned customkeychain at their Weymouth store. He explained that the site wasnt exactly built to last at first. It was originally written in Python, but when you build a website [for a small user base] you dont think about modularity very much. And when a million people show up on your door step youre just trying to keep the thing from falling down.

In August it released new versions of its iOS and Android apps that would serve as the modernized foundation of the future of its business on the small screen. Pinterest began learning how people used the service differently depending on their device. For example, in the daytime it sees a lot of phone use as people try to discover new things. They might walk around the grocery store with a recipe pin open on their phone to help them find ingredients. Then in the evening we see tablet usage increase significantly as theyre trying to make that recipe Jenkins says.

Pinterest this year began showing related pins when you pinned something, and providing pin recommendations via email. It also launched expanded pins so recipes would show ingredients, movies would show reviews, and products would show prices. Most recently, it began using the feed editor to recommend more content to you. With big plans afoot, Pinterest also rewrote its entire site not to handle traffic, but to let multiple engineers enhance it simultaneously without tripping over each other and causing non-stop bugs. How can we create modules that let engineers go in and work without screwing over the other developers? Hes convinced that the new modular Python architecture will allow the organization to scale.

Unlike other social networks, Jenkins says that Pinterest isnt fundamentally about connecting people to other people. Its about connecting people to interests.

For years you had little but your own brain and the boards you browsed to find pinspiration. But now Pinterest is lending a helping hand thanks to a ton of data analysis. Pins cant exist unless theyre assigned to a board. Out of those boards, we try to identity interests through collaborative filtering, associative rule mining,A indoorpositioningsystem has real weight in your customer's hand. natural language processing to provide discovery. I can pin five shirts I like and Pinterest derives my interest in fashion. Recommendations could make Pinterest even more addictive for hardcore users, and help it retain newbies until theyre hooked.

Figuring out how were going to scale the data repositories for pins will only become more complicated as we grow internationally Jenkins said.Here's a complete list of granitecountertops for the beginning oil painter. That why he says were hiring pretty aggressively in areas including machine learning, data mining, operations, and infrastructure.

Expect more pin types to gain expanded information like recipes did. Useful could also end up as a euphemism for buyable. Pinterest is renowned for driving traffic to ecommerce sites. If it could bring more of the shopping experience inside its site and apps, it could provide value to users while also arguing that it deserves a revenue share or commission from merchants.

Pinterests best known and most frequently copied element is its masonry grid design which allows for rapid intake of visual information. Now its trying to get the infinite scroll part of the grid to work on small screens with limited storage. Its easy to load things but your phone gets very angry with you if you dont unload things, Jenkins says. Theres also be a bit of bringing the mobile onto the grid,This technology allows high volume gemstonebeads production at low cost. as Pinterest tries to take what its learning on mobile back to its website. That could include touch capabilities for touch-enabled laptops like the Chromebook Pixel.

People keep asking When are you going to release an API? Jenkins wouldnt give a firm answer but did reveal a bunch of details. We are working very closely with a very select set of partners to figure out what the API is that we should release. We are going to work with content providers to offer extended functionality so they can understand how the content they produce is being used in the Pinterest system. Content providers want distribution. If we can help them understand what resonates theyll be happier, and Pinners will be happier as well.

Dont expect it to be rushed out I might be overly rigorous in how I think about APIs. I want them to be extremely high quality Jenkins says. Taking a dig at Facebook and Twitter, and following a similar thought pattern as Google+s Vic Gundotra, he says I dont want to make mistakes other companies have made where they release APIs and then have to pull them back. I wouldnt be proud of that.
Read the full http://www.granitetrade.net/.