Wednesday, September 28, 2011

What are you waiting for?

So, said the reporter, "Excuse me... What's your opinion on the meat shortage?" to a nearby standing clique which had an American, a Russian, a Chinese and an Israeli. Now, the ability to perceive the reporter's question was preceded with the idioms and metaphors of the culture and economic context of the individuals who were from their respective countries, as usual.

- "What's a Shortage?" queried the American. In the land of excess this was an alien word.

- "What's meat?" quizzed the Russian, for, he had not heard of any for a long time.

- "What's an Opinion?" wondered the Chinese. As much as what had been the rule of the land.

- "What's an Excuse?" inquired the Israeli. That's something they were never used to !

Now, there was an Indian passing by who happened to overhear the Reporters interrogation. *The Indian understood the question perfectly well*. 

The point is? Well we Indians as startups and Innovators happens to be in the context & economy that allows us to be blessed to understand all the connotations of that question. Yet, we must hope that it would be a blessing, the day we stop understanding "excuses". This shall probably be the day we accelerate the process of Innovation.

Whats your execuse to not start-up? or to not push the boundaries? or to not stop whining about lack of eco-system? or...

(Attribution: this context is a adaptaion of Mike Leigh's Jewish play "Two Thousand Years")

Monday, September 26, 2011

Mobile Apps Developer? Ad Monetization? Really?

There are so many mobile app developers nowadays. It is getting hard to keep track of what exactly each of us do. Whenever you veer into the Business and Revenue model discussions with some of them (us), you get to hear awesome things about Angry Birds! The pitch is to follow the Angry Birds revenue model. What is that you might wonder? Oh the Ad Monetization based Revenue models you are pitched.

But have you explored what App you would like to develop? What is the consumer and target groups you are going after? The utilities you enable for the target group? About other revenue opportunities? As soon as these questions flow, it sometimes look like I catapulted a Black-Bird that explodes on contact at them, and the conversations flow like, Oh, we 'thought' about it, we have tried it 'once'! those are pretty 'hard' things to do!

Thought about it? Tried it Once? Hard to do? I am gasping for fresh air at this moment.

You know, there is absolutely nothing wrong in aspiring on a good Ad monetization revenue model (even if you are not benchmarking Angry Birds here). I know several great Ad monetization businesses which are doing decent top-lines (some of them are good friends and well respected). So, before you break it down, its important to know what other Business Models exists.

At the very high level, the following are some broad brush strokes mobile app business models :

  • Enable mobile channel play for existing web channels which are non-ads based (Deals, Events, News).
  • Enable mobile channel play for Business LOBs creating new media asset (Leads, Branding, Merchandising).
  • Own a focused, curated content (created through your app), and charge for the access to assets (POI, Metrics, API).
  • Enablement of mobile mediation layers controlling access to existing enterprise services. (Enterprise Mobility).
  • Mobile device management and control platforms. (Enterprise Device Management).
  • Develop apps for other business as a service. (Software Services).
  • Develop compelling utility/game apps which users are willing to pay (Pay per download).
  • Develop compelling utility/game which are free (Ad monetization).

The Revenue models can be categorized as follows:

Not charging the consumer but charging a Business: (Free for consumer)

  • Charge for Action (Number of leads filled, Deals closed, Events booked, Close loop).
  • Charge for Impression, Click Through Rates (Brand content disseminated, Click on assortments).
  • Charge for Service to develop (Fixed Cost, Time & Material, Hybrid).
  • Charge the Developer or Enterprise for Usage (Charge for API calls, charge for throttled use of mediation layers).
  • Earn through Promotions (Ad monetization). (This is diff from 'charging' for Impression where you control the sourcing)

Directly charge a consumer:

  • Charge for Commodity (One time download fee)
  • Subscription Charge for Access (Subscription for monthly news, Stock tickers, Domain content)
  • The above can have a variant of free for limited features, different monthly fee for different capabilities.

Since we started on Angry Birds Ad monetization as benchmark (for a specific Business/Revenue model combinantion), you should know some high-level metrics of Angry Birds (as I write) if you are aspiring to be one. Angry Birds have in excess of 350 Million downloads (across all platforms) and has 200 Million minutes of play time everyday (that's equal to 380 years of game time every day). Also, it is not just one game anymore. It is more a game platform with series of games for every season (Halloween, Christmas),  along with extension packs and Themes for movies (Rio) and Ports (Magic for Nokia). Angry Birds is rumored to have taken in excess of 50+ release attempts before it struck the exact game psychology with its adopters. They also have a paid version, where 12+ Million copies have been purchased on iOS.

Then, Of-course, you were kidding when you mentioned Angry Birds. Yeah, I know that. in that case, lets focus on the people who are making a decent living off Ad-Monetization, shall we. Lets take an analogy here, there are so many bloggers who are making a decent living off their blogs cause they are experts in what they write and have a focused audience. So when you say you would like to start a blog to earn money, be very focused on the audience. Even for mobile apps, audience and utility matters. Who are your audience? demography (age profile..)? psychography (likes, tastes...)? What app are you developing (Game, Lifestyle, Info, Travel, Shopping)? What is the experience you are targeting (Lean-Back, Lean-Forward) etc.. Once you have figured this, you need to worry about:

  • Hardware, Software, Device (and experience) fragmentation.
  • Compelling UX, Addictive Visuals, Theatrical Experience.
  • Content Refresh Rates (or game levels).
  • Challenges and Puzzles for the right psyche and device profile.
  • Context Awareness (Location Based, Event based...)
  • Seasonal, Theme or Memorabilia (Christmas, Disney, Harry Potter)

Next, you have to worry about the metrics of the mediation layer. In other words who shall give me ads? what shall I earn? when do i get paid etc.. Some of the Ad and usage metrics that you have to be aware of are as follows:

  • Dwell Time: The average time spent by the user daily.
  • Frequency: The number of times the consumer uses the app.
  • Time lags: The time between frequent visits (this can taper down).
  • Shelf Life (Half-life): The time that a app can keep a consumer excited, once a half-life is reached, the usage tappers down.
  • Impressions: Number of times a ad is served.
  • CPM: Cost per Mille of the ad impression.
  • Clicks: Number of people clicking on a served impression.
  • CTR: Click through rates, number of clicks by the number of impressions served.
  • Global Fill Rates: This is a percentage metric representing the inventory of ads that is served.
  • Funnel Fill Ratio: Number of new users, acquired users, median users, half-lifed users that are in your funnel.
  • Viral Coefficient: This is a fancy ratio, but in general this means, how many of your current users referenced new users.
  • Relevancy Metric: Is it in native language? Is it specific to a locale? is it specific to a demography? psychography?

Notice that you earn less money by serving impressions and more for covering clicks (clicks are considered as actions and are equivalent to leads) So, how do you increase your impressions and clicks? This is exactly where the rubber meets the road friends. If you do not know your audience profile, or track the usage metrics of your app, then, you are in complete doldrums here. You need to have a good analytical dashboard in the backend or rely on the one that your ad mediator provide (like InMobi or AdMobs).

Assuming you have a compelling App which has a addictive user Interface you 'still' need to ponder on:

  • Is this a lean-back app driving dwell times but is not so much focused about frequency?
  • Is it a lean-forward app relying much on recency and frequency?
  • Do you uderstand the events that produce the time lags?
  • How about your Relevance strategy? What are your ad promotion strategy based on such relevancy?
  • Do you serve text ads versus graphics?
  • What about the Global Fill Rates? Are the house ads (free ads) eating up your ad inventory?
  • Who controls the fill rates (you or the mediator)?
  • What is the seasonal cost of such inventory (Oh, these rates fluctuates beyond anyone’s understanding).

Just while you start interacting and getting a handle of some of the above, your half-life hits and usage tapers down, now you start worrying about Funnel fills, refresh rates & median CTRs for retention (through new versions, new levels, extension packs). What about virality? marketing budget to go after prime adopters? Yeah, I can go on, but by now you get the point right. If you are throwing more than one app out there to spread the bets on the portfolio, then you have to do all of the above for every app if you need a decent top-line. But then again, you can blindly just throw apps out there not worrying of anything and hope to make some revenues as well (Good luck if you are one of them).

Just in case if you think you are in the Ad monetization business because it is 'Easy'. Think through again. Talk to established players and understand their metrics. Get a good product mentor on board. Most importantly be ready for a wild experience. Oh do not forget to have fun in the process. Cheers :)

Thursday, September 22, 2011

Social Graph: The Philosophy of Existence.

The first post of this social-graph series explained the scheme of things and the second post looked at the broad types of activities one can perform over the graph and the experiences that are possible using edge devices.

In this post, I ponder (philosophically) on the nature and the context of the social graph from the perspective of why it matters or why people participate.

Its been found that even bacteria-infecting viruses (called phages) do communicate through chemical markers (viral gene expressions based on genomes and proteins) in the host system to make collective decisions to either remain in a latent state or to attack the host. The point is? Well, the point is that life, it seems, is inherently and implicitly social, both by nature and context (voluntarily and involuntarily).

In-fact, i have a amateur-theory that all organisms are hardwired by nature to engage with other organisms in order to get fitter in the cycle of evolution. The fitness is based on the 'access' and evolution of 'shared' knowledge (of a particular gene state in a host, or the economic output of a country, or any other esoteric inter-galactic protocol).

Given that humans are organisms blessed with superior interaction skills, our incentives are primarily driven by engaging with each other, often emphatically, based on many different 'intents of life'. The intents of life can be shaped up from a functional context as understood from previous posts or also based on 'tastes' and 'interests'. Note that I am generalizing all intentions that forms different context and still call them as social-graphs . In many places, people have coined many other terms such as interest-graph or taste-graphs and have provided interesting connotations for these topologies (I will provide my opinions on this in future posts). For me, simply put, if Individuals are socially networking, then its a social-graph. The rest of it is a signalling discussion around the context. IMO, the intents only bind the domain of discourse of the graph establishing a context.

To illustrate, If I am interested in bird watching, I would love to engage with fellow bird watchers to exchange notes, discuss, and learn about this domain. Now, will I choose to search for others sharing similar interest over my friends network (Facebook, Twitter etc..), or will I form a separate group, or will I move to a separate social-graph for bird watchers is again a discussion of semantics. Lets for a moment assume that there is a bird-watchers fan-page or a group as part of a established social network (say Facebook) and call it a platform.

Every time people engage with each other over a topic of interest over a domain (on a platform), they refine the knowledge-base within that domain and take it to the next level (evolution). Note that in the process of engagement, I may also want to date, hang-out and party with people who are like minded. The pleasure modules in the brain is incentivized in this 'interest seeking' behavior to feel good about engaging with other humans on the topic of interest that are close to our heart(?). In effect, when we say we engage socially, we involve the faculties of both Cognition (understand, learn) and Psyche (perceive, emote) to derive conclusions of such engagements. it is almost a driver of life, that we voluntarily or in-voluntarily participate in social-graphs then.

Given the connectedness of the digital world, it became quickly obvious that establishing context specific platforms over the Internet (the digital nervous system) to enable such interactions is a natural hit. As hypothesized, it becomes natural for us humans to gravitate towards such platforms which enables access to others so as to share our life's intent enriching our experiences.

The nature of such engagements happens over different styles of communication as follows:

  • Synchronous or Asynchronous (relevant based on the flow of info and immediacy) 
  • Closed or Open ended (status versus questions)
  • Private or Public (Inside a closed virtual room or on a wall)
  • Unicasted or Broadcasted (one-to-one or group messages)
  • Organized or Un-organized (hashed, tagged, categorized)
  • Structured or Loose (Text heavy or template based)
  • Animated or Bland (Exchange of Narratives, Theatricals and Videos)
  • Location specific or Global (My locality, state, country...)
  • Discovery versus Recovery (Find something new versus fetch something I know)

Again, it has to be noted that the styles of engagements are chosen primarily based on the combinatorics of dimensions that we have understood in the past posts such as the functional domains, the edge-device profiles, the device experiences (upright, lean-back, lean-forward) and also on the intent of accessing or sharing information and the conclusions that are sought for such engagements. So, again, not all engagement styles are suitable for a given combination of device-profile, device-experience, environments and engagement intents. (much to ponder and innovate here BTW)

Now, then, it starts getting interesting to study and analyze the usage profiles and behavior of humans (to access and share info) through such bounded context of different social-graphs. The intents that are captured, the knowledge disseminated, diffused, built or contained in such graphs becomes extremely important to push the bounds of human evolution a bit further. Of-course such knowledge gets extremely enticing to all practitioners from the functional domain, be it marketers, sellers, buyers, recruiters or also to linguists, sociologists, politicians, economists, logicians, biologists etc...

We can conclude that Social graphs do matter and it is natutral for people to participate. And, what about the evolution? The reach and relevance, unprecedented. The potential and imagination, unbounded.

Wednesday, September 21, 2011

Social Graph : Access, Activities, Environment & Experiences

In continuation to the previous post : The Scheme of things which may have set the base to understand the usage of terms such as Identity, Relation, Context, Functional Domains, Functions, State, Reputation and Edge Devices, I intend to work towards the functions of the Access points (Edge devices), Nature of Activities and Environment of use of the social-graph in this post.

Edge Devices
To quickly summarize, edge devices are the access points or tools that are used to communicate with the digital world, be it world wide web or the social graphs. Broadly there are 3 types of edge devices today being PCs/Laptops, Tablets and Mobile. Each device has advantages and disadvantages around form factor, resolution, input functions (keyboard), nearness (always with me?), bandwidth, processing capacity, memory etc.

Nature of Activities
Lets now broadly categorize the 'nature' of activities that a user performs over a social graph, as follows:

  1. Heavy text content creation/updates (Initiation as well as operational)
  2. Media content updates
  3. Passive Information Consumption
  4. Active Information Consumption
  5. Interaction

Heavy text content creation: During the initiation (or the first time use of the graph) there may be Identity creation, profile updates and other such social-graph-context specific info that needs to be updated. Depending on the context of the graph, these activities may include extraction and load process (Ex:Bill-of-material) into the graph as well. Also on a ongoing basis, you may change, update, add several stateful information that is relevant to the context.

Media content updates: These can be updates of Audio, Video, Pics or other relevant media specific to context.

Passive info consumption: User may visit the social graph site based on events like a notification (friend posted something on the wall) to only consume that portion of information. Or, the user may in general check the news-feed to be in-touch with activities and statues of others through a quick browse.

Active info consumption: User actively searches for a specific profile in the hope of hire (Linkedin), or a product search to find the seller (eBay) etc. There is a active and intense engagement of user to retrieve info either in the form of discovery (looking for something) or recovery (recall what i know already exists). This requires users attention to perform the activity with satisfaction.

Interaction: Interaction may involve, chatting, tagging, commenting, likes, purchase, filling lead forms, quotations etc... (diff function across diff domains of social graph)

The most important aspect of the access points are its environment.

- PCs/laptops are typically fixed point devices used from home or work. They are fully loaded across all capabilities including larger screen, ergonomic keyboard, powerful CPU, memory, fat-pipe (bandwidth) etc. Users typically sit-upright with full attention to perform a task when working with PC. There is limited distraction and ambient noise when user works on a PC. Lets call this the sit-upright posture. Lets say the combination of the device capability and the posture is called as a experience. Typically the consumer can perform all activities (#1-#5) with absolute ease during this experience.

- Next, Tablets are mobility devices meant to be carried around and used in environments such as conference, work, shop-floor, home, events etc. The device capabilities are typically built around info consumption with a focus towards, touch inputs, relative mid sized screens (compared to PCs) and a good processing power for gaming. Now these devices are built to be used in a relaxed manner, leaning-back on a sofa, reading a digital book, playing angry birds, taking quick short notes etc. Lets call this the lean-back posture. IMO the functions that are mostly apt for a lean-back experience is media-content-updates, passive info consumption, and some amount of active info consumption, and interactions. Not all functions of active info consumption may provide maximum utility for a lean-back experience. Also heavy text inputs or extraction-load processes are a complete no.

- Finally, Mobile phones are extreme mobility devices which have the smallest form factor, limited screen size, not so powerful processor etc. The most important aspect of this device is its always-on, always-with-me profile. Also the environments in which they are used are far wider in range than the above two device profiles. Lets hypothesize that a user who has a PC + Phone, would mostly switch to PC when in home or work. (I will keep the discussion on users who never have a PC but only a phone, separate from this post). This eliminates home/work as the major usage time of phone as a access point to engage with social graphs sites. Given this assumption, phones, almost overlap with all environments that of Tablets.

Mobiles are also majorly used in shopping areas, airports, cafes. Typically due to the limited form factor of the mobile, and the amount of ambient noise, distraction, grab-of-attention etc that happens to engage the user away from the phone, the utility of the phone is primarily built to be able to be operated with minimal attention span of the user and maybe single handedly. User typically leans-foward, quickly consumes info and gets engaged back in his environment. So lets call this the Lean-forward experience.

Lets extend the context of the mobile phone profiles a bit. There are feature phones (Not so powerful, limited capability devices) and there are smart phones (A mini Tablet lets say). So the utility of each of this differs based on input capabilities as well as info consumption (based on processing power). Assuming that mobiles are used in a highly distracting environments (unless in a cafe), the nature of activities that you may perform on a mobile gets further limited to a quick status update (about where you are, what you are doing) or passive info consumption. Also only some amount of media content updates and interaction can be performed. Performing Active info seek, or heavy operational text inputs are almost bad experience from the utility-value-grid perspective here.

So the users usage behavior on accessing specific functions of the social-graph is completely dependent on the profile capabilities of the Edge-device, Environment, and the Posture (Upright, Lean-Back, Lean-forward). In effect, all these dimensions combined, drives the experience of the user which limits the maximum utility of a social-graph function for a specific combination of these dimensions.

We can conclude that the way we experience the social graph significantly differs based on the combinatorics as explained.

Here-in lies a premise to focus on either incremental-innovation to build on top of the permissible utilities of a specific experience (mobile-lean-forward) or work towards disruptive innovation on making the pain-points go away (speech-to-text for heavy input processing). That said, it is difficult to change the environment of use or the limited attention a user gets for an experience such as mobile-lean-forward.

Will leave you for now on the thoughts of how applications and domain specific social-graphs can evolve particularly around a mobile-lean-forward experience, as majority of the graph utilities are yet to emerge in this segment...

Social Graph : The Scheme of Things

I was off-late wondering on the nature of social graphs in general and what maybe the functions and context of use of these graphs. Social graphs being the root of many business and scientific discussions, requires a common scheme of things that are used in well defined way in the conversations to broadly make sense.

I am breaking down my thoughts into series of posts to manage the size.

In this series, I thought I can take a first-cut of my understanding and opinions on what this scheme can be (in a common business like language) and represent a preamble to my future posts based on social graphs.

Probably the following distinct scheme of things are obvious at a higher level during social graph discussions.

  1. Individuals who participate in a social graph. (Identity)
  2. Relationships between the individuals. (Relation)
  3. The Domain and its functions. (Context)
  4. The standing of an Individual within a social graph based on the values, behavior and outcomes of his activities. (State, Reputation)
  5. The tools that are used to communicate. (Edge Device)

Lets get the academic definitions out of the way

Social: refers to the interaction of organisms with other organisms and to their collective co-existence, irrespective of whether they are aware of it or not, and irrespective of whether the interaction is voluntary or involuntary

Social Graph: A graph that is constructed by connecting the organisms to each other forming relationships and 'How' they are related.

Defining each of these schemes further

Identity: Each individual will have to establish a unique ID to distinguish herself across other within a graph. This can be email, TwitterID, FacebookID etc..

Relation: The nature or type of relationship you hold in real life, such as, employee, employer, friend, spouse, buyer, seller etc..

Context (Functional Domain): A social graph can be formed between humans relative to many functional domains. Which means, it does not necessarily have to be between a type of relationship such as friends (friendship being a functional domain). You can construct a social graph of organized sellers in a market place relatively connecting with each other to create the sell side dynamics like price control, logistics, bill-of-material, turns ratio, availability etc. You can also have a social graph between organized buyers in a market place connecting with each other enabling the buy side dynamics. Then another graph having participants from a work hierarchy (linkedin), another between a seller & a (adhoc) buyer in a market place (eBay et al). It can be between structured sellers (Amazon, eCom...), it can be for a specific community of aggregators (who are not producers or consumers) such as co-op societies. You get the idea of functional domain specific social-graphs by now.

Functions: Note that the relationship drawn between two people connected in a social graph has certain functions that are enabled to make sense within the context of that domain. In Facebook for example, you chat between friends, you are more casual in your status, you poke friends, upload family photos and comment on them etc... But when you get to Linkedin, you are more focused on getting connected to people from your industry, hiring, marketing, selling, defining job responsibilities accurately, enable recommendations etc. These are the functions you perform over the domain specific graph to engage meaningfully.

State: You seamlessly cut across all the functional domains of different social-graph that you belong to. You as a individual know your personal state and verify your state against activities across different graphs. You also update your residual state which is very specific to the functional domain of the graph. For sake of brevity, each graph carries only enough context and state to play within the domain boundaries of the graph. To illustrate, you may update Facebook that you just became a father (and your demographic profile in FB). You may update your new promotion and brief job responsibilities on Linkedin. As a Individual, you are aware of both these states, but as LinkedIn or FB the graph only knows what it needs to know.

Reputation: Based on your state, activities and interactions in the social graph, you manage to hold a reputation as an Individual within the functional domain. In Facebook maybe you are known to be overtly quite, but extremely chatty when it comes to Twitter. You may be a super connector when it comes to LinkedIn, a top commentator when it comes to Disqus. Reputations are based on likes, rating (and other esoteric parameters such as relationship weights based on frequency, recency, reputation of people you have connected to etc... PeerIndex and Klout are examples to manage such reputation). Also graphs themselves internally manage reputation based on internal algorithms which may be used to show you news-feeds of friends who you interact the most etc.

Edge Devices: There are many types of devices that can be used to access social networks today. PCs, Tablets and Mobiles being the broad category. Each device has advantages and disadvantages around form factor, resolution, input functions (keyboard), nearness (always with me?), bandwidth, processing capacity, memory etc. Assuming there is connectivity for the device to access the graph, the functions of the graph that you would access is based on the context of the graph and the capability of the device to allow a reasonable experience for the consumer to engage with the graph.

Given this scheme of things, it probably becomes meaningful to have reasonable discussion about a social graph in any functional domain without getting overtly confused on the terms used. In the next post I will talk more on the edge devices and its functions which sets the base to understand the pivots of utilities when the edge device capability amplifies or hinders a social graph function.