Collecting Thoughts

Put up this post on medium a couple of weeks ago. The idea was to capture some concepts that could be useful to mobile app marketers, PMs, and developers - sort of like a 101 on how to look at your app’s metrics.

Maybe I should have titled this post as "Subsidizing your new phone connection using Uber". I typically don't activate international roaming on my mobile phone. I get a local number while on travel. I spent 60$ on a mobile connection in the US earlier this year, and $36 more recently in Toronto. On both occasions, I realized that I could more than make up for the money spent by sharing a Uber referral code with my new mobile number, thus collecting a free $30 worth ride on the new account, and another $30 on the old account that gave me my referral code. In fact, one could even do the same with multiple cab apps in that city and get more free rides. This is probably a huge arbitrage opportunity in India where it is really cheap to obtain a new connection, though the credits are typically low at around Rs.300 (~5$). Now I wonder if the value of referral promo is determined by Uber as roughly 2x of Cost of obtaining a new mobile connection. This way, it at least wouldn't make sense to game the system and go after obtaining more new connections to earn credits from these cab apps. Of course this is tough to replicate in the Indian market where new mobile connections cost close to nothing.

My fascination with Mobile Messaging started in 2010 when I started thinking deeply about building a asynchronous voice-based group messaging app. I remember briefly discarding my dream when Facebook bought Beluga in 2011.

But then was always the itch that no one was doing the voice messaging part right. A year and few months later, the itch grew stronger again, and discussions with my friends Vignesh and Deepak lead to a little more belief around the potential. We could revolutionize how people converse! Potentially make conversations more personal and meaningful - especially with that larger circle of friends you don't find the time or excuse to call.

We tried out the experience of some of the existing apps in the space. We saw news of Voxer raising $30M. We saw a few more entering the space. Checking them out made us feel asynchronous voice has real potential, while still feeling only those who experience it can understand the power. We had to get people to try out voice messaging to get them hooked.

We quit our jobs in 4Q 2012 to start building  (I had to help out remotely at Jigsee for a little longer as there was an acquisition that was happening). We were encouraged to read articles touting some of the new folks in the space like YC-backed voicegem as potentially "the next Instagram" (billion $ exit - which sounds low now in the space? ;) ).

We studied how KakaoTalk, LINE and WeChat are big in their geographies, and set out to build PhonOn as our own unique messaging experience. We launched PhonOn for iOS & Android on Feb 14 2013. We introduced a pro-active approach to prompt users (ala Telegram does now) to reach out to their friends who joined the platform or to wish friends in voice on their birthday and make it personal.

It worked! Our users started conversations from our prompts - in fact 90% conversations were in voice. Those who used it loved the experience. But alas, we had depended on Facebook for our events and prompts, a platform that by then was viewed as spammy by most in the circles we had managed to seed our app in. A lot of these messages were going into blackholes as the other party never saw the message.

We eventually fixed the experience/invite mechanism, but by then it had become harder to explain why someone should try PhonOn. WeChat and LINE had entered our home market India with a big bang - doing television advertising during the prime time programming. Bharti's hike was doing well too, with the ability to incentivise users to try their youthful app. To top it, Whatsapp too did a retake on their voice messaging (I'd say more as a reaction to WeChat and Line). Of course, FB had launched the same a lot earlier.

We had been the only ones carrying an experience that put Voice and Text on an equal footing in the messaging experience (till Whatsapp did their update to bring the same experience). We were also unique in that our voice experience was engineered to work well even in 2G networks and lower end android devices. We also had 2 unique messaging units in picture messages captioned in voice and "talking stickers" (fancy animated emoticons that would talk or have your voice added to them).

In spite of this, it was increasingly getting tough to convince folks to try our product. It looked like we either needed a big telco backing us, or needed to turn this into a content play of sorts (with the "talking stickers"). But the market was getting crowded every day and we decided not to go at it on our own and focus on another problem we had discovered along the way.

We had seen the new trends emerging of mobile messaging getting popular within companies - as a way of employees engaging with each other within the enterprise. Altassian had a product, microsoft had a product, and apparently folks like dropbox, box, etc are building this too.

It was clear that that light weight asynchronous nature of messaging was too good for folks to resist in any environment. So why is it only people engaging with friends and families or within the enterprise? Why not businesses and brands engage with their customers through the same medium? This thought led us to work on Konotor. We had built the platform as a way for us to engage with our own app users on marketing our new content, introducing our app, and conversing with our users on feedback or support.

We spun it out into an SDK and took this to a few apps, who immediately lapped it up! The idea of reaching out to groups of users through a simple messaging experience and allowing them to engage back was a strong value proposition to a marketer/product manager, as well as the support teams. We could see this especially hold good for commerce oriented apps who needed the ability to do personalized and contextual messaging with their users. Our SDK, segmentation tools, APIs and conversation experience worked to help re-engage and retain app users and give them confidence to transact in these apps.

We now touch over 3 million users, around 170K of whom have explicitly engaged back with a business or brand, and many more have implicitly consumed the messaging via Konotor powered inboxes. We're happy with where we have gotten since January this year - a successful pilot with Target, many good names on board, winning Qprize India 2014 - but there is a lot more to do to make this a success.

I've had my share of start-up ideas before I finally took the plunge 2 years ago. Some of them had names and domain names even, while others were just "ideas"

Here's a few of them:

Pensieve Online

This was to be a way to collect your memories and share them selectively at different points in time with friends/family. The digital version of the "Pensieve" we know from Harry Potter! I gave up www.pensieveonline.com last year.

I believe with smartphones, google glass-style gadgets, actively tracking and recording moments from your life for posterity has become even easier. Maybe the right time for a digital pensieve is still in the future!

Granting access to memories for short periods of time is also easy integrating with your social networks or phonebook.

At one point in time I was hoping Friday from Dexetra would sort of go this direction.

Deal Gully

Deal gully was to be a "map-driven" deal search. Me and a friend (Piyush) came up with this as we watched Groupon grow mighty fast in the US and many clones pop-up world over. Piyush might still be paying for dealgully.com for all I know :).

We built the aggregator that would scrape all Indian deal sites for the deal metadata, and then added map locations to them (manual in some cases).

The core idea was to allow you to filter by a category of interest, and only see deals in the boundary that matters to you (instead of all deals in a city).  You could then subscribe to the combination of categories and geo-fence of sorts.

[We dropped the idea when we saw that the Indian deal sites did not show you deals that you would regret missing. The same deals would keep popping up on all the sites - which meant there was no urgency or excitement in availing the deal.]

Online Get-together + Serendipity (No Name)

The difference between meeting your old school friends at School or a common hangout vs. doing a meeting online is usually more about how some of the serendipitous experiences that you have in real world disappear when you try to do something online.

The idea here was to let you do your online meetings with picking a location (say your usual hangout). You can individually have your visibility and "eavesdropping mode" enabled/disabled to let other groups meeting in the same location see or hear you. That way, you can bump into others you know or even strangers, and have some context to interact with them.

--

3 weeks ago, I participated in my first hackathon - organized by Capital One in San Francisco. I really enjoyed the experience and would definitely do it again :)

The focus of the hackathon was to build for small businesses. One of the 4 major problem areas identified was "customer engagement". We were told small businesses in US spend 5 times more money to acquire new customers than to retain customers, and hence customer support and engagement was very important to their business.

For those of you who don't know - I build Konotor, a mobile-first user engagement platform, focused on re-imagining how marketing and CRM works in the mobile world. We provide an SDK for app developers and work with some of the top mobile apps in the commerce space in India today. So naturally the customer engagement problem appealed to me and my team set about creating a solution for this.

Here's a gist of our hack for the event:

* We built a small business owner a simple mobile app to publish their deals/ announcements/ etc through simply sending a text or picture message (Over The Top - like whatsapp/iMessage). This is something they loved and appreciated for its simplicity and familiarity when we showed them the experience.

* A mobile app through which a user could follow small businesses. This may be a local coffee shop or a mom and pop store or any local service provider the user usually deals with. Now the user can (a) chat with the business for support or queries by simply messaging them from the app - text, voice, picture messages and (b) get notified of updates/deals offered by the business of their interest when they open the app

There is no private messaging platform today connecting small businesses with their customers. We thought of the one app for all small businesses vs. auto-creating an app for each small business and went with the first option for the hack (not far from what my friends at Haptik build, but for a different target market).

I have since shared this with a few small business owners, and they love the simplicity of the idea, and also the option of having their own app they can give out to their loyal users as a way to provide support, tips and tricks, announcements, etc. They aren't all on twitter and feel there is too much happening there to be noticed, but can see themselves paying and using this product to stay connected with their loyal users.

Answer by Srikrishnan Ganesan:

Lets look at some metrics - installs to downloads ratio, active user cohort analysis, sessions per active user to make some inferences: 1. 1300 installs from 3000 downloads seems low. How do you count installs (or do you mean unique users who ever opened the app?)? Are you incentivising downloads in some way? One would expect upwards of 80% of downloads to open the app at least once if its organic. 2. 300 active users per week in itself is a meaningless figure to look at. Given it is 23% of your installs, it looks like a decent number - but could be bad if its all new users in the last week. What you really need to measure is what % of users from the last week came back this week, the week after that, etc (retention). There is a lot you can look at in terms of active users (splitting new and repeat, looking at ratios of weekly/monthly active users, daily/monthly active users, etc). These help you benchmark your app’s performance in terms of engagement and retention. Still be careful to look at these when your reach is larger than now.   3. Sessions per active user in a week also is an indication of engagement. 7 sessions per user seems good for a week. Watch out for (a) outliers skewing your info at this stage (like your own usage of your app might be so large it skews the sessions count total), and (b) accuracy of counting sessions (can be tricky/incorrect on android if done via onResume in an activity, etc) You should also look at time spent per user on average.

By Srikrishnan Ganesan

Can’t believe this has taken till 2013 November to happen (writing this note a month after it happened):

https://developers.google.com/app-indexing/

A year ago I was thinking of building a deep search for apps (knowing there would one day be a standard for crawling and indexing apps). I was debating with friends if companies would be open to offering their in-app data in a structured format, for indexing by a third party search tool, in exchange for more traffic/downloads sent their way.

I guess the answer is yes in general. Apps want to be found. The only concern being opening up data for indexing (through a standard mechanism) also means a whole host of competitors crawl your data on the pretext of also being a search engine. Happened on the web, will happen on mobile once a standard evolves. So far, mobile only apps haven’t had to worry about bots crawling and stealing their content.

Google’s initiative is a starting point in this direction. I think it caters more to the big guys who are on web AND mobile, and would rather take the user to the app experience from the search results. Not sure it is yet meant to help discovery of mobile-only apps.

Justdial (NSE:JUSTDIAL) makes its revenues from local businesses in India who advertise their services through JustDial's voice-call driven business directory services. Apparently their revenues were ~22% of Google's revenues in India last year. Users often find it convenient and accessible to call JustDial for contact info of a local business, rather than "Googling" for it. Plus there is the matter of having an updated directory from their "feet-on-street" operations.

I use a Nexus 4, which was updated to the latest OS version Android v4.4, codenamed KitKat. The new Phone application (dialer/contacts) on this update makes it ever so easy to find and call a business, as long as your device is connected to the internet. Just punch in the name of a service you are looking for, and you start seeing results from your city offering the service - as though they are on your phone book! Tap on a result to call.

Now, I find this convenient! A percentage of my searches on JustDial are going to move to Google. I might call sometimes since I expect the JustDial database to be fresh and more complete. Having worked on local business listings collection during my stint at Rediff.com, I can appreciate how JustDial might still have an edge. But I believe this update should be a very big deal for JustDial, and I see some others share the sentiment.

There is also another angle to local search, stemming from the hypothesis that its not just listings, but knowing what businesses are trusted by your friends that helps you make a wise decision. I can see Google's "+1"s coming into play here at some point (maybe right within the dialer experience), and there are also interesting start-ups like Frilp that seem to be taking on this challenge.

The JustDial app feels too cumbersome when compared to the Google instant style experience (search as you type) provided on the Phone app. Neither has JustDial has done a great job of social yet - most people don't have their social networks formed on JustDial. Two big problems for JustDial to figure out as more people come online from their mobile phones in India.

(Note: DBC/ "Death By Chocolate" is a popular ice cream sundae offered at the Corner House Ice Cream outlets in Bangalore).

What should JustDial do to avoid "Death By KitKat"?

Over the last 10 years, I've made decisions that have had big impact on my career and life. While I don't ponder much on "what if?", I have given some thought around how I make these decisions, or simply, "what is my decision making framework?"

I recently came across this video where Jeff Bezos talks about his decision making framework - his optimization function is "minimization of regret".

This led to me looking up online on what are some of the common decision functions, to see if I am of a particular type, at least with respect to career, business, etc. There are the usual suspects in there - maximizing upside, minimizing downside, maximizing potential upside (probabilities factored in), and so on. Good place to start: wiki.

I couldn't slot myself in one of these directly. I am a risk averse person (yes, I don't think entrepreneurship is "risky" for me - because I can always get a job again if I need to, and because of the strings).

My decision making framework is something like this - "What's the worst that can happen from here? If I am OK with the worst outcome, I'll go ahead for whatever has better upside."

I mostly don't look at the probability of the outcomes at making these decisions - if the chance for success is low, and chance for failure is high, that's fine as long as the failure is one I can handle. However, I don't consider the very very low probability (in my mind) outcomes (for the downside or upside).

You needed to be in the BBM group to really be "part of the gang". BBM was a rage. This was the true even as recently as 2011-12 in many city colleges in India.

When BBM does reach iOS and Android devices around the world, I will probably give this app a shot again. Why?

Privacy - You don't need to worry if random people, service providers, who might have your number saved to their device see your status updates and DP changes

Multi-screen - If BBM comes out for the tablet devices as well, that makes it more attractive than some of the messaging apps that are restricted to devices with a phone number

I miss all the banter we used to have in the days of BBM groups. Can't say if it disappeared with people getting off BB, or if my "gang" just grew older, and decided not to message as much (work, and non-QWERTY devices).

Scanning through many early stage apps and their reviews on Google Play, I was surprised that users have been demanding features or fixes from app developers in this manner: "Do XYZ, and I'll change my one-star rating to five stars!"

Why do these users resort to unpleasant tactics? Would they do the same if you were a friend?

I recently saw some tweets around MAU(monthly active users) milestones and remembered this article by Suhail Doshi.

If someone shares MAU without retention numbers, I can only get more suspicious about their churn. Why? If a product has buzz around it (which in itself is of course an achievement), it is going to have a high rate of sign-ups, and the new additions can contribute significantly to the MAU. That doesn't tell you the story of what happened to your additions from the previous months (how many did you keep?).