AppSurfer

AppSurfer in its current form is trying to solve a problem of unnecessary app downloads, by making apps instantly accessible. We are trying to do that by streaming apps from the cloud and running them instantly on user’s devices. Recently Google launched a new product called ’ Instant Apps’ which solves the same problem but with a better approach at the platform level.

Looking at this scenario, we had a long, hard look at the potential options for our product/technology. We see that the problem we had set out to solve has been solved in a big way, and we can’t compete with it going forward. So we have decided to call it a day on AppSurfer and shut it down.

We have spent the better part of last 4 years building AppSurfer. The journey has taught us a great deal and this is all about the journey of building AppSurfer.

Where it all Started

We were a group of friends, who happened to be engineers and had the startup bug. We were in android since a long time, and we saw the android was going to be a huge deal in mobile space. This was 2011, still early days for android adoption. We wanted to liberate android from devices, put it on the cloud and use it from browsers. We thought the possibilities were endless if we could do that. We used to call it the ‘DroidCloud’.

In 2012, we bootstrapped and created a prototype of AppSurfer - a technology that could run android on the cloud and stream it to browsers. It would allow users to access apps through web browsers from any device. We presented it in UnPlugged and Pune Connect and got a lot of feedback in terms of market positioning. We also got into ET Now’s TV show called ‘Super Angels’ which was great and geared towards funding new startups. We got into its finals and got funded by Vijay Shekhar Sharma from One97 - the company behind Paytm. We also got Alok Kejriwal as an advisor for us.

Version 1.0: ‘Try before Download’ platform for apps

In its first iteration, we created a ‘Try before Download’ platform for consumers and an online app demoing platform for developers.

For consumers, we created a web and mobile based portal where they could try apps without downloading and if they like it, they could install it. We presented the platform to the Indian audience at the Mobisparks and Demo India, and the buzz around it was positive.

We courted developers by providing them free app demoing widgets, which could be embedded all across the web. We also created ‘Try Now’ buttons that they could put in web pages.

We launched this platform in early 2013 and got a great response. We got covered several times then forth on TechCrunch and within months, we had more than 8,000 Android developers, uploading close to 40k apps on our platform. We were serving millions of app views and did more than 500,000 app demos.

But then, we had to figure out how to monetize the platform. Our plan was to monetize with ‘App Install ads’ in consumer platform, but sourcing problem with top apps made it very difficult to hold a consumer and engage him regularly.

Version 2.0: Monetization by offering App Demoing tools for Developers

We started offering app demoing tools like App widgets, Try Now Buttons and App Pages for developers to demo the apps all across the web, inside Facebook NewsFeed and drive the downloads. We started offering ‘tier based’ pricing for using the demoing tools. During this period, we got a chance to showcase the product at TechCrunch Disrupt 2013 and the US audience loved the offering.

We had a small revenue from the developers, but it was not big enough to continue in that model. We needed to pivot into something where we could maximize the value of the technology.

The positives out of the use case were, the rate at which people were going for downloads after doing the app trial was really good. We decided to use this insight and use it at a place where we could monetize it in a big way.

Version 3.0: ‘Try and Install’ ad format for App advertisers and Ad networks

We created a new product, a new pitch, built an entirely new technology stack suited for the product. We launched the Android app in the US and Europe in beta. The response was overwhelming and our value proposition hypothesis was proved.

But by this time, we were almost out of money. And we were not able to raise additional funds at that time. So in early 2014, just before the startup boom in India, we ran out of money.

Staying in the Game

In case you didn’t know, having no money is not good for anybody, let alone a startup under a pivot. We had very few options in front of us. We had a small core team, a product that needed a lot of technology updates to be suited for a new model, and no money. We believed in our product and future potential. So we decided to buckle up and raise some cash ourselves. We started doing consultancy for other companies and building up cash. We moonlighted on our product and kept on improving it.

The last Effort Version 4.0: Instant Apps without downloading

Last year, we had built up enough cash to launch our product in a new model. We saw a big opportunity in making apps instantly usable without unnecessarily keeping them downloaded on the device all the time. The way we did it is by streaming apps from the cloud Android and running them instantly on user’s devices.

We created a new product, a new pitch, built an entirely new technology stack suited for the product. We launched the Android app in the US and Europe in beta. The response was overwhelming and our value proposition hypothesis was proved.

We also presented the app again in the US at major events, this time, it was DroidCon SF and the media loved it. Everyone needed it and thought it could be improved.

Recent Google IO 2016 saw the introduction of ‘Android Instant Apps’.  They solved the exact same problem with a better approach at the platform level. We think, having better control over the platform, ecosystem, and mobile devices helped Google to come up with this approach. The way Google does it is by splitting an app into modules and allowing the user to download only the module which they wish for at the moment. This makes the download size small and makes apps instantly usable. And we think this is the best way Instant Apps could have been brought to the users and only Google could have done it by introducing the concept at the platform level.

How AppSurfer Instant Apps work

How Google’s Instant Apps work

The Way Forward

A startup lifecycle is full of ups and downs. This is exactly what AppSurfer went through and this has been a great journey. Appsurfer got introduced into the market 4 years ago.We built a great team, moved fast and became first movers in the space. The journey saw AppSurfer move through various pivots and had its business model changed various times, raised funds, had great investors backing it and created a patented IP.

We tried real hard in last four years, spent everything we had on the product and company. We had a great time, learned a lot, knew our way forward and kept going. Two of our products eventually got built by Google itself, and now people world-over will be using them.

Today we are shutting it down, with a sense of pride in the effort and sadness at the result.

Thank you to everyone who has supported us over the years.

It has been roughly two weeks from the first user instance running on AppSurfer cloud until today and we have learnt a lot, received overwhelming acceptance from the users and beta testers alike. We are ready with our second version of Beta app and here is what all we have included.

For the first timers here, AppSurfer is a platform which is changing the way we are using mobile apps today by removing the need to download an app before using it. This is a very small change in the overall flow of using an app but it changes everything for the end user. Now, using AppSurfer, anyone can reach any page inside an app #Instantly without having to download the entire app.

It is similar to reaching a destination web page inside a website without having to download the whole website. Downloading the whole website? Seems outright ridiculous, right? And so is downloading an app to make a payment or to book a cab or to book a hotel or to eat out at a fancy restaurant. We simply don't do these things that often to have these apps permanently installed on our phones anyway. And thus, we have AppSurfer.

Head here to know more about the AppSurfer App

AppSurfer Beta 1: Achievements and Shortcomings

Since our release 2 weeks back we have clocked over 6074 total app server sessions (Apps used inside AppSurfer) with an average usage of 3 minutes. These figures are pretty overwhelming considering the fact that we were expecting a modest figure of a thousand users to register within this period which was easily overthrown by a 3938 registered user figure currently. And there are over 12,000 users signed up to become beta tester and waiting to get white-listed for using the app in the US and Europe alone.

We have a robust beta tester community including early adopters, influencers and mentors who have been regular and constructive in their feedback to get the app refinement in tune with user behavior and needs. Some of the crucial feedback we got include:

- The streaming quality being slower in some areas.

- Frame rates being lower than 30 FPS

- Responsiveness being slower than native app

AppSurfer Beta 2: The improvements

AppSurfer Beta 2 is set to remove all the shortcomings of the first Beta release with main improvements in streaming quality and responsiveness. The key upgrades are:

- We have brought the FPS to 30 with auto-bandwidth streaming.

- Improved event streaming with QUIC protocol.

- Improved resolution and DPI for a clearer, cleaner and native-like experience.

This is not all, we are introducing Intent Based App Search. Which means anyone can now search an app, not by its name, but, by its intent. This is something we noticed from the user behavior. People, more than often, search for 'booking a movie ticket' rather than 'Fandango'. This will improve the overall experience for the user with him reaching his desired app with relative ease and would also help in exploring new and exciting app for any given need.

Droidcon is not only an event but also a religion in itself with the App Universe uniting to know what best is happening in the app world. AppSurfer also attended Droidcon SF and had a lot of fun. Presented our offering of Instant Access to apps and met some really cool people in the process.

Thanks for visiting us at #DroidconSF @appsurfer! Here’s your awesome spin powered by @googlecloud pic.twitter.com/NGmtRS6dEB

— Google Cloud Spin (@googlecloudspin) March 17, 2016

Droidcon is a perfect stage to launch a tech-intensive product as AppSurfer, considering that Droidcon is invariably packed with some of the smartest brains in the business. The response at Droidcon was a little overwhelming and more a confirmation of our thought processes behind the app. People do need a solution which makes it easy for them to access all the Million plus apps in the world and they need an escape from having to download them to use it.

Almost everyone we met was keen on the things this app will stand to change. Consider the fact that you would save on storage space, time and processing power using this app. This is huge and people want it now. Here is an interview of AppSurfer team from Droidcon by Expedition STEM, which shows the positivity and enthusiasm we received at the event.

Also, this year the focus was on solving some of the key problems in the app world today. The keynote by Chet Haasey and Romain Guy from Google into the current world of Android was informative. Congratulations to all the speakers at the event including Will Bailey from Facebook, Ty Smith from Twitter and Dan Lew from Trello.

We are getting an overwhelming response from the initial users of AppSurfer Beta App and we are more than just happy.

One critical feedback which we got more than often was to enable ‘Search for Apps with Intent’. People are using AppSurfer for searching and accessing apps in the Android world for fulfilling their intent.

Today, we are releasing improved App Search on AppSurfer.This would enable the users to not only search an app with the name but also to search an app with the intent, like, booking a cab or booking a movie ticket or renting out a car or just recharging your mobile phone account.

The below video showcases the new updated inside the new AppSurfer App.

There is an ocean of mobile apps in the world. On average, a usual user has 40 apps on her phone and uses a very handful of 10 apps on a regular basis. These figures don't add up to a great exploration of million plus apps in the world. While there are a lot of reasons for not using an app, we just didn't want 'download' to be a reason for it. Downloading an app is a cumbersome process and exactly what we are trying to remove.

We have been working, for a long time, on building a system which would eliminate the need of downloads for using an app. We wanted the users to have the seamless experience of Instant Access for an app, any app, they wanted to use. We have finally achieved it with AppSurfer. AppSurfer removes downloads from the experience of using an app.

We launched AppSurfer App here in the US this week. It is a beta release and we need beta testers to test the app. Here is the play store link to the AppSurfer Beta App: http://bit.ly/1RSzLG1

AppSurfer is all about removing 'Installs'

The experience of using an app for the first time is a long process.You have to search the app in the store, download it, install it and then launch it.

It is a terrible user experience for anyone to be forced to download and install an app that they don't use on a regular basis. No one really needs or wants to download an app just for a payment or one-time story or an offer. This consumes a lot of time, bandwidth and patience from the user end. It also brings in an additional effort from the user to manage all those downloaded apps to optimize your storage.

I attended GopherConIndia this weekend. On day 1, there was a talk by - @beyang on Meta Programming in Go. He showed various examples about using parser package to parse go source code and write simple analysis or generation tools with it. I found it very interesting, and wanted to try it out.

On afternoon of day 2, I decided to give it a try. I have been writing few simple web apps and apis with Go for last 5-6 months. Especially for Json/REST apis, I had to write a lot of json tags for Struct fields. Because we have fields written in camel case, and we want json fields to be in snake casing generally. So I decided to write a small utility that will take go source as input, search for structs and add tags in snake casing.

I finished up with a small utility of around 100-150 lines. Once I shared it on twitter and reddit, I got some useful feedback about the implementation. Especially this one - https://www.reddit.com/r/golang/comments/2wo9xy/easytags_easy_jsonxml_tag_generation_tool_for/costdis pointed out that I was unnecessarily using a lot of string manipulation, and that I should look at format.Node. So finally I changed things up, and now this tool uses parser, ast, and format packages of standard library to good effect. I would love to have more suggestions to improve the code.

Gopheregala 2015 was one of the most amazing hackathon we have participated in, till date. Golang is the first love of every web app developer in RainingClouds. Being one of the early adopters of Golang, We were very excited about Gophergala and decided to co-host the event in Pune. 

Golang (incubated in Google) is gaining a lot of popularity among developers. Just to gauge the interest among the people in Pune, we had co-organized a "Golang Primer" event. We have been developing some amazing products for our clients in Golang for the last one year, so we decided to share our experience with the amazing and interactive crowd we had for the event. Here is our deck for the event. We met a lot of interesting developers keen on shifting onto Golang. We were glad to help them out to clear up some of their doubts.

One of the odd things for Gophergala was, it was overlapping with a long weekend in India. Despite of that we saw a great support from the Pune gophers. There were 4 teams coding live from Synerzip office.  

We also had participated in hackathon and ended up with an automated API documentation creator middlewares for the web apps written in Golang. It was a great experience for us. Our first version was ready in almost 9 hrs with documentation. We have got a good feedback from the community as well, 23 stars and 2 forks in 5 days :). Here is our screencast for YAAG : Yet Another APIdoc Generator.

Here is GitHub repo of YAAG. We have tried to give support to as many web frameworks and libraries as we can for YAAG. And we are committing to continue the development of YAAG in future to make it a full fledged document maker for Golang web apps. 

    Monday, 1st Sep 2014, the day on which I joined, one of the reputed startups, RainingClouds Technologies. Coming from background of IT giant, I was extremely excited. 

    On first day, I saw working environment and I was amazed. Work culture is very much similar to startup cultures from California. Also the bean bags, I don't use them frequently but it helps a lot while writing a complex code. Moving on, I was aware of shockers which I will have to face.

    The very first shock I faced was instant and very fast software development. Yeah, It was a shock for a guy like me who is from lazy IT giant. I started writing code after 3 hours of joining company. Yeah 3 hours! Which would have taken almost 3 months to start a development after a client call in other MNC giants. Never mind, I was in a startup! Where you can not afford to do mistakes. I had to work on large projects and take initiatives, work hard and write code, and lots of code. But it was fun to write code while listening to music. This kind of freedom you can get only in a startup. I never got such freedom in my previous company. TCS it was, relax finally I have revealed my previous company name.

    The parents! Yeah, I had to convince them to join startup. As they knew after joining a startup I may not get time for family and I really will have to work hard, may be on weekends, late night. And all these they thought came to my destiny within 4 months. Unfortunately I had to work on weekends and late nights, but fortunately I experienced all these. As it was an unique experience. 

    I always wanted to join a startup, so when I did, everyone keeps asking me "How are you doing? Does your company offers you as much holidays as mine does?" And many more... I just smile at them. That elaborates my answer every time. It does not matter for how much holidays I get as far as I am growing. Frankly speaking my learning curve is perpendicular to earth. Just kidding, wanted to express how much I have learned in such less time.

We had launched app demoing tools 1.5 years back. Over this period we have continuously improved the app demoing feature and tried to make it grow. We have received 8000+ android developers signups across the globe and ran half a million app demos till now. But looking at the data and trends, we have finally decided to turn off the app demoing feature.

If you have publisher account, what this means to you -

  App demoing tools i.e. app widgets, promotional pages and feedback pages will stop working from 15th of February for free users.

For paid users, service will be operational till the end of billing cycle or 15th February, whichever is later.

All publisher accounts will be deactivated on 15th February 2014.

We regret the inconvenience caused to you with the closure of the app demoing feature.

We have built an unique technology stack and it has many compelling usages. We have seen promising results with ‘Interactive app advertising’ offering and it would be going to continue. Also we are cooking something interesting and will be launching it shortly.

Demonstrating an app with localized setting (geography, language etc) is consistently requested by our developer base. We value it and so rolling out a feature to enable demoing of app in local languages.

Now apps can be demonstrated to the users in localized environment of device. You set a locale for your app and AppSurfer will run the app using the provided locale.

Here are the ways to set locale for you app - 

  Setting Locale for App -

For all app runs that happen, this locale will be used.

For setting  app locale, you just have to pass locale as specified in app APIs here - https://github.com/RainingClouds/AppSurferApi/blob/master/sections/apps.md#2-create-app

Locale can be set from app create API call or app update API call.

* Currently app specific locale can be set via API only.

  Setting Locale for Widget -

Locale can be set for particular widgets by adding `locale` parameter in widget iframe URL. For e.g. to have 'fr_CA' as locale, you can add `locale=fr_CA` to widget iframe URL.

Locale is set for widget as well as app, widget specific locale will take priority.

Widget API also provide a way to get the widgets with specific locale set. Check API documentation here - https://github.com/RainingClouds/AppSurferApi/blob/master/sections/widgets.md#get-widget-code--

  With Javascript API - 

Javascript API for AppSurfer widgets are coming soon. It will let you set locale, get logs for currently running and app much more. Keep an eye on blog for release news :)

GPU support has been a long standing request from our developer community. We have worked hard to get it integrated with our platform, and get it right. We are now proud to release, GPU support for OpenGl 2.0 based applications.

We have come up with an innovative solution to optimize the performance with GPU and keep the costs in check. Having GPU backed apps, makes the whole thing that much more resource intensive. Hence, we have built a dual system. We have a direct GPU support, backed by AWS G2 instances for 3D games with high performance needs. In addition we have a cost effective graphics support with a 'Software Renderer'. This keeps the application costs low while keeping an optimum performance.

The support is automatically enabled for all existing applications. If you have an OpenGl2.0 based app, now is a good time to put it on AppSurfer.

Notes:

Time and again we have been irritated by the way ads are presented in mobile apps. Static content like screenshots, banners and one way communication like video are the core issues in these ads. Humans prefer content that is interactive and immersive. We respond better if we are involved in the action. But currently mobile app install ads are neither interactive nor engaging. Do you know that 88% of people don't find ads as engaging and continuously decreasing ROI in ad market reflects the same.

This is now about to change. We have come up with a brand new mobile app install ad unit - <a>Try and Install ads</a>. Now mobile app itself can become an ad. Users can play with the app inside that ad itself, and if they like it, they install it. Neat, isn’t it.

And what’s more, it works with all the current ad networks out of box. It is basically a Rich Media Interstitial ad unit which is supported by all major ad networks.

Now how do we know, that this thing works as prescribed? Thousands of developers have been using AppSurfer tools (App-widget, App pages, Embedding inside FB for demoing apps for over a year. After running about half a million app-runs we have found that, the tools are performing exceptionally well. We have seen an average 16% CTR (Click to download) on our tools on web. The most promising area where we could apply this tool is mobile app advertisements. Here engagement and interactiveness is a real pain for app advertisers. So we decided to build a brand new ad format called 'Try and Install' unit. It lets users instantly ‘try an app’ before going for download. It’s highly engaging and interactive way to show what you are selling.

We have tested it across our mobile properties and seen that our thinking was bang on. We have seen exceptionally well CTR i.e 16% which is 7X better than normal ads.

One more important point that came across was app retention rate. As the user has already used the app before installing, the app retention rate becomes very high, and the advertiser get very high life time value from each install. So these ads not only convert with a higher rate, they also give you quality and targeted installs.

What is 'Try and Install ad'?

It's HTML5 interactive ad unit, which let users 'Play' with the app instantly before installing it on device. It's highly engaging and interactive form of ad unit. You can use it as a standard Rich Media Interstitial ad unit.

How does it look?

You can try the live demo with AppSurfer here: http://www.appsurfer.com/interactive-ads

How does it work?

a. Once you submit an app to the AppSurfer. AppSurfer creates and hosts interactive ad for your app.

b. In return, you get a creative ad URL for an app, which you can use in any ad campaign across ad networks of your choice.

Key Points: 1. 7X higher CTR for HTML5 Interstitial Ads. 2. Nearly 4x increase in app retention rate. 3. Compliant with MRAID, supported by all major ad networks.

<a>Get in touch</a> with us to start using this ad unit.

Reference Links:

We have put our efforts in making more and more apps compatible with AppSurfer. We have supported Google Maps V1, native code , accelerometer so that more and more apps in different categories become compatible with our platform. All this while we have added incremental performance updates over the last one year. Now we had reached stage where we needed a breakthrough performance improvement for the app widgets. 

So we set out to rethink entire graphics delivery stack with these goals -

Apps should run with minimum of 10 FPS performance, on 800kbps + connections.

Minimize the bandwidth consumed in by the widgets.

While doing this, we devised a few approaches to achieve these goals. 

We decreased streaming graphics update quality for fast changing screens.

  We sensed the screen stability, and once screen stabilizes we increased the graphics quality.

                 Intelligent graphics streaming would be needed to minimize bandwidth consumed. Hence, we built the system to stream only when there are updates available on screen. Also, when a screen changes we only stream the diffs, hence bandwidth remains low.

We primarily engage with android app developers and related ecosystem. Time and again, we have had requests for customization's, tweaks, and little changes. But as a product, we have our own plans and roadmap.

We believe that ‘Android on the Cloud’ can have very many use cases and multiple markets. We want entrepreneurs and innovators out there to explore these opportunities. Many existing businesses can leverage our technology for their specific needs.

So now, we are opening up our platform and our unique, ‘Cloud based Android’ stack to the world. We are launching our own API, through which you could leverage our technology and platform, to build your own innovative ideas.

Head to AppSurfer API for more details. See Programr for a live integration of AppSurfer’s API. Here is API documentation.

Having worked in a number of companies over the years, my urge for knowing about UI/UX has come and stopped at an ironical thought. "The more i know about UI/UX the more i realise that i know so little". To keep learning more and to keep myself on my toes I couldn't let go of an opportunity to work in a startup. The first day itself I was handed the most important aspect of AppSurfer - the App Trial Widget. 

<Since then, I have changed the App Trial widget, learning to make it better with each version. And yet I am far from perfect. Before i came to AppSurfer the existing widget then(version 1), as shown was limited and far from polished. The first thought that struck me was it was far from being a real phone. It was a nice mockup of an android phone but still a mockup. So I came up with version 2 for the App Trial widget as shown.

The web and design releases were fast paced and a startup's stringent deadlines made me roll out a widget with minimal tweaks and maximum UX. I searched for the latest Android phones, and keeping the widget inline with the Google Nexus I designed the phone to mimic it. Users could easily identify a Google Nexus and the phone looked more real.  The previous widget had a 700*700 dimension, which ate up a lot of vital page space. To accommodate for better space utilisation i chipped the widget from the centre and kept the top and bottom curvatures as the previous Nexus had.

The rotation, GPS icons were placed inside the widget in version 1, making the Widget look less phone like. The grey background and old iconography of the icons made them difficult to understand. In this second version, I shipped these icons out of the widget on a white background giving them more visiblity. The iconography used was the one which users could easily relate to.

Recently I reinvented the widget. This time I have chosen a flat design for the widget to suit our future release of AppSurfer. Our latest widget is inspired by the Note 2 design. The power and sound buttons on the side are included to get closer to a real phone. The rotation, GPS icons are better placed along side the widget rather than hanging in no man's land. This placement of icons make them strongly bound with the widget and easily understandable. The play icon is changed to "Click to Run", to give a feel that the widget is to be interacted with. The play icon used before gave users a sense of a video. The white color makes the widget look fresh. For widgets that are placed on other blogs and pages, we are going to release an alternative black widget. 

Last 2 weeks have been really good. We have improvised on our technology constantly and now our infrastructure costs have come down drastically. This has opened a number of usecases for us to tap on.  Based on user feedback from you guys we have launched a developer centric solution.

Developers can now avail our full set of Pre-Launch demoing and App Marketing tools for their apps. Our experiments with App Discovery have enabled us to design these set of effective tools for devs. In this stiff App Market devs grapple to gain more visibility for our apps. Keeping this in mind we have designed these tools, which affect your pockets marginally.

Our new release includes

-       Embeddable App Widget – As before devs can use this widget to let users try their app. Customize the widget based on your app page suitability. As we always say let your marketing material talk for itself.

-       Customizable App Page - Our app page got 10k hits in last 2 months and we generated 2k leads to downloads. We converted 20% on our App Page. On similar lines we have created a customisable app page for your app which you can enable in a click. This page is optimised for mobile, tablet, desktop browsers. Have a look at a sample app page.

-        Feedback Page - We have enabled a link based or privacy based feedback/testing page. Your trusted users can have a look at your App Feedback page and drop in valuable insights on your app pre/post release. Have a look at a sample feedback page.

-       Play Inside Facebook - Directly engage the user on Facebook. When your app is shared on Facebook, it can be tried in the facebook newsfeed itself. See this on facebook

-       Universal Compatibility - We support html5, flash. We support app runs on all devices ipads, tablets, desktops, smartphones etc.