0

Classification Vs Clustering

I invariably get confused between the two. More because of the Classication of living beings, animals , insects, human beings etc we have been taught in our primary science/biology classes. Clustering , i have been always seeing from the marketing perspective of segmentation. Does both implies grouping in one form or the other? Absolutely not.

Say we have a population given, when we find similar groups within this population, we basically divide the population into clusters. The similarity is on basis of few characteristics/parameters say age , disease which help us to determine whether one is young, middle-aged or old. Before the start of the clustering process , we dont know which characteristics will be deterministic helping us to find the DIFFERENT SIMILAR GROUPS. There might be few outliers left which will not have any aligning characteristics to any such group.

Classification already have the grouping done. We just need to determine the group in which one particular element of the population will fall into. We start with a training set of data which already have groups mentioned. Depending on this dataset, we find out the pattern or the group/class function which will help us determine given a new population , which class the members of this population will lie.

The below figure stands apt for classification problems. We just need to decide whether a given book will fall into the algorithms box or science box or philosophy one. We categorize the books into known classes. These classes/groups and the number of them is already known (Supervised learning).


While clustering , we simply have all the books at our disposal and depending on their contents we identify what are the various groups/class of books in the library depending on which we label the box in the rack. The number of groups/classes is unknown when we start (Unsupervised learning).

The other example can be news item - whether a piece of article falls into the Economy news or Marketing News or Technological news category et all. This is classification. Finding all the categories under which news item will fall is clustering. Deciding whether a mail is a spam or not is classification while grouping the mails into different labels one of which can be a spam label is clustering. 

So is clustering always done before classification?
0

Would I Go on a Date with You?

Now this is an interesting usage of Analytics : Data-based deep insights , observations and facilitation on the dating world. What if you answer the few questions in the chart above and get to know if we are the right match or not?What if you have a similar map of yours and we able to match our compatibility?  Want to know the best questions to ask on your first date - those easy to bring upfront , those not so personal and obvious ones (few of those on the lower right quadrant below) yet handy enough to know what you both really care about?


Want to know your date's political , social or religious inclination in more subtle acceptable way. OkCupid ,  does it all for you.

How it works
It encourage you to ask a series of rather interesting questions and generates a match percentage (they have applied a patent for their algorithm) depending on below inputs:
  • The answer
  • The expected answer from one's partner for the same question
  • The degree of importance one gives to the question asked, weightage given in bracket
    • Irrelevant(0)
    • Little important(1)
    • Somewhat important(10)
    • Very important(50)
    • Mandatory(250)
  • Whether one's answers can be made public or not 
Suppose there are two questions with following responses(the green and blue color denotes the match) :


Importance Weightage/Answer/Expected Answer
A
10/1/1
50/1/3
B
1/1/2
250/1/1

Q1.
Q2.

For B:
Total Point Available from 2 questions=∑Importance given by A= 10+50=60
Answer to Q1 only matched, so points gained=10
A’s Satisfaction level from B=Points gained by B/Available Points to B=10/60=16.67%

For A:
Total Point Available from 2 questions=∑Importance given by B= 1+250=251
Answer to Q2 only matched, so points gained=250
B’s Satisfaction level from A=Points gained by A/Available Points to A=250/251=99.6%

Match Percentage=sqrt(16.67%*99.6%)=40.74%

In the quest of getting more and more matches , one ends up answering as many questions as possible (there are tons of them) and hence gets generated the map of one's expected matching criteria along with the valuable insights. The response to whether one's answers can be made public or not gives the ability to segregate the question database into less private and more private as depicted in the second figure above.
0

Do you have a Second Life?

An INSEAD Campus , a running economy of its own with its own currency (Linden Dollars) , a Currency Exchange and GDP, a free market where you can buy or sell virtual goods and services,entertainment, a profitable real estate biz (you can buy and sell virtual lands and make a profit in Linden Dollars and finally get it converted to REAL ACTUAL non-virtual US Dollar : 1 US$ = around 260 L$) , Festivals and Celebrations, boyfriends/girlfriends or escorts for hiring , Banks , ATMs and shopping centres et all - Second Life from Linden Lab has it all.

More than 60 Fortune 500 companies have established their presence here. Coke has set up its vending machines and selling millions of cans of virtual drinks (just think why these virtual drinks are being bought !). Corporates are using it for business meetings, education/training and simulations. Having no need to buy food to survive, several residents here are making real-life money from SL businesses making it a commercial hub.


Your involvement just depends on how deep in the virtual world you want to get into. And remember it's not another so called Massively Multiplayer Online Role Playing "Game" wherein you face some challenge and try to overcome it. One common aspect is that you are here to enjoy just like any other game. Still , if you are a fan of a Quake, Doom, Halo, World of Warcraft or Age of Empire, your chances of adopting the SL is higher. You being a resident have a perceptual immersive experience with entirely new set of capabilities(eg: you can fly) in this virtual Linden World. Create, design and offer products and services limited by your own imaginations. Slowly , it is turning out to be the online media for the firms to do pilot testing for new products , experiment , sell and promote their brands.

How Linden Lab makes money 
  • Advertisements from Second Life viewer/console and Marketplace where goods and services can be bought and sold for the virtual world.
  • Gets commissions for items sold in the marketplace.
  • Currency Exchange: Transaction cost is charged when users buy or sell L$ with each other.
  • Land Rentals make up around 80% of the revenue. The users can buy or purchase public or private piece of virtual land for a monthly fee.
  • Premium Accounts : These are tiered subscriptions which gives user a certain amount of land.
But as we know, our online behavior can be very different from our real life one and so there is still a question mark over how successful any consumer/market research activity will be on this platform.Plus,there is no government rule but still you cannot behave or cause harm to the experience of others as there are establised measures placed to tackle such behaviour. The fact that the richest 10% of the users control around 90% of the money in this economy makes it less appealing to competitive users.

The transition from early adopters to followers is still a long way ahead. But still , a thriving economy Second Life has become, it will indeed be interesting to watch how it will evolve and facilitate the next level of innovation by breaking the geographical barrier at least virtually. And don't get surprised the next time you hear of a Sony Island where they have a virtual headquarter with a window office !

Product Management Challenges - II

This article is continuation of earlier Product Management Challenges - I and highlights the tough road being traversed by this genre of managers.

Competitive Landscape in an “IT Doesn’t Matter” era
If we go by Nicholas Carr’s famous article, Information Technology as a whole is becoming a commodity which no more would give a sustainable competitive advantage. So , business acceptance of a software product for purchase considerations is slowly getting apprehensive. “Why to invest further on IT if they are able to go business-as-usual?” or “Today we are implementing Oracle Apps R12, tomorrow we need to migrate to Fusion, there is no end to it”. So any software product should provide ample value proposition to the customer which compels him/her to try/buy it. Add the competitors to it and your product sustainability and customer lock-ins are always at a threat. That’s why we see many product start-ups initially working in stealth mode, doing their own internal pilot testing , getting the relevant product and consumer insights to get a differentiatied offering which would outsmart their competitors.
Resource Crunch
This is especially in the area of user experience designing. There are very few visual and interaction designer and many times their scarcity limits the firm capability to form the crux of the product team. If there are no designers working side by side with the product managers & engineers, it will limit what each of them can do.

Consumerisation
In layman’s view , it refers to the blurring line between the consumer and enterprise technology needs. And many a times, a new technological product is launched first to ordinary end user and once the product gets popular amongst the masses, it is being offered to the enterprise customers with additional value layers. For example, Google’s Application products were first opened to the masses and then it came up with the enterprise version. Apple recently launched Joint Venture, a tech support offering for small businesses covering the products like Mac, iPhone, iPad -- the very same ones which got sold off to general consumer. The challenge is to formulate your product strategy and roadmap across these blurring lines , to accept that the B2B and B2C distinction of end product usage is fast getting extinct.

The Successful Company Trap
Seeing the Google making hoards of money and awed by the fantastic products line from Apple, are you thinking to do product management like them? Be cautious. The point to be noted is that while Google & Apple have their own models of product management, the same might not give you a success formula. Their model is fit for them (until their founders remain so deeply involved in their product offering).

Cloud : SaaP vs SaaS
Cloud is both a challenge as well as an opportunity. If the existing product line is not web-based , then the firm has the danger of losing out on its existing and prospective customers. This is particularly important in emerging markets where the consumer is not ready to pay too much for the software product and hence the rise of ad-supported freemium business model in these markets. With SaaS the switching costs are decreasing plus the anytime anywhere access of the product which brings newer challenges to the product manager to continually improvise on the product for customer retention.

Customer vs Consumer
Often the people that make the purchase decision are very different than those that actually use and and are dependant on the product. The product manager ends up being responsible for a product that will be sold to (or used by) people he has never met, to people that speak a language he doesn’t understand, subject to laws that he isn’t aware of, and pay by mechanisms he has never used. The platform product management is a good example - Adobe offers Photoshop which is being used by slew of designers to create media/flash applications which are finally ‘consumed’ by totally set of people.

Product Management Challenges - I

Coming in early , working relentlessly until late nights , day after day , week after week but still not satisfied and happy ; packed days with back-to-back meetings , burst of emails piling up , the changing customer dynamics, the race to catch up with emerging trends & technologies and the various stakeholders’ pressure for that one magic product expected out of your team – Welcome to the world of Product Management !

The thrill of getting a successful product being used by the whole world comes with its own set of challenges to be faced by the product managers.  This article discusses a few of these challenges:

Lack of Role Clarity
Great software products come from the collaboration of product management, product design, engineering, and marketing team. One important aspect to the organizational design is that these teams should be equals but invariably we see one getting buried within another. The organisations have a tendency to mix/overlap the Product Management with the Product Marketing roles and responsibilities. Hence , many times we might see Product Managers striving to create their own niche space and acceptance within the organization, proving subconsciously that yes they are required and needed. My short product management stint made me notice a sort of humorous role confusion : The business team thinking that the product management guys are from the engineering side while the engineering team having a notion that they are from the business side ! So the role clarity itself becomes a challenge.

In a nutshell , we can have a very high level distinction:

Product Management
Product Marketing
Responsible for defining – in detail – the product that the engineering team will build.
Responsible for telling the world about this product
Defines Product Requirements
Defines Marketing Requirements
Activities include Product discovery, idea generation , ensuring that the product is useful, usable and feasible,  optimizing the end-user experience.
Activities involve : Pricing, Positioning and messaging, Promotions, Launching, Online marketing , Customer Acquisition Strategies.
Both should collaborate effectively for the product to succeed

Stakeholder Management , Buy-ins , trade-offs and Consensus Tendency
One of the biggest challenges is convincing others like the top management about any innovation's utility. Selling the product idea and concept to the internal customers who will fund the project for product development is not an easy nut to crack. Lack of required upfront support and funding increases the time-to-market.

A very effective way the product discovery team can handle this scenario is by creating a prototype which can be shown to the stakeholders. The nature of the discussion will be more productive when stakeholders can see a clickable prototype versus something just talked about in air or in some form of paper/powerpoint specification.

Within the organization , Product Managers need to aliase/interact with many parties : product marketeers , project managers, interaction designers, visual designers, usability engineers, prototypers, engineers, architects, QAs , cross-functional teams, etc. The decisions are being taken in concensus and become more of a group activity. The product objective shufts from coming up with something great, to coming up with something that keeps everyone happy and doesn’t get you fired which actually negatively impacts the product innovation. Seth Godin said: “Nothing is what happens when everyone has to agree.” While consensus is something nice to have, we basically need more of collaboration wherein the teams can work together closely and provide a fusion of customer value/functionality, usability, and technology.

High Expectations
Yes, the customers would expect your product to solve all of their problems and your CEO would want to roll out the new release at a breathnecking speed. And if you are an engineer plus MBA breed , don’t get surprised if you are being considered superhuman. The best way to tackle is to communicate a very realistic picture to your internal and external customers. If the expected deadline is too tight , break the work to be done in chunks and modules ,  identify the actual achievable roadmap and convey the same assertively. Using numbers, some sort of data crunching to prove your point helps invariably.

Product’s Success Metric
There is a lack on clarity on how exactly a Product Manager’s performance can be articulated except the fact that it naturally should be linked to the Product’s Success. But again , there is no best way to measure a product’s success. Should it be through the revenue generated? Or Profit? Or the number of end-users? All of these parameters can be useful but they don’t really provide the exact picture.

One simple business metric coming up recently (and adopted by the likes of Apple, eBay and Amazon ) is called the Net Promoter Score (NPS). The customers are asked how likely they would recommend the  product, on a scale of 0 to 10. Those rating 9 or 10 are called “promoters” (they’re out there telling others to adopt the product); those rating 7 or 8 are called lukewarm or neutral; and those giving 0-6 are called the “detractors,” which are not likely to recommend the product to others, and may even be actively warning against the product.

NPS = The percentage of promoters - The percentage of detractors 

Balance between Domain and Technological Expertise
Product Management requires skills like assessing opportunities, defining product principles, product discovery, and prototype testing. The whole idea is to discover a product-market fit and this requires a deep understanding of both the market needs and the technology’s capabilities. The challenge is how to keep pace with the domain/market trends and rapidly changing technlogies when the job requires your day-in and day-out effort. This is very critical while you interact with various teams and plan for product/feature feasibility. Even the product lifecycle is also getting shorter putting on you those extra worries to speed up in order to monetarily benefit from the existing product as fast as possible.
0

The Next Level of Indian BPO

Remember the Chetan Bhagat's One Night @ The Call Centre ? Do you think that Indian BPOs still work the same way - Taking calls and that's all?? Millions of so called easy jobs , mostly youth , good number of women employees, coffee, video games, chats, nights shifts - there is more to it.

There is a paradigm shift happening in this industry. In fact the global BPO industry is striving hard to strike through the growing manpower needs emerging out of new opportunities here. In India, there is a lot happening. The industry is no more limited to handling voice calls; the protectionists sentiments and economic conditions in western world are playing their roles; slew of mergers and acquisitions carving out cultural coherence; the 'cloud' buzz is finally taking over the BPO landscape as well. 

Voice to Non-voice
Yes , the later one is gaining momentum and growing faster than the voice-based sector. Its sheer common sense. Just look around on the web, better say social web. We now have customer services handled through twitter feeds; there is so much being spoken about and communicated to the firms through non-voice medium. The logic is simple, it costs me to make a call while the costs on web communication just go unnoticed, the minimal(commoditized?) it is. Then there are niche areas from healthcare to procurement to engineering services to advanced analytics - the part of KPO stream which actually never required voice-based support. And believe me, none of them are low-end works nor what we call as client's non-core part of the business. We are getting it less on the traditional cost arbitrage advantage and more on the value add and expertise plus time zone advantage our part of the world is able to provide - call it Intellectual Arbitrage. You can see several new gen indian startups catering to each of these niche non-voice based outsourced services. PharmARC Analytics caters to pharmaceutical & healthcare sectors; AutoQuill in Architecture, Engineering and Construction (AEC) domain; Elixer Web Solutions and Datamatics in Recruitment Process Outsourcing.

BPO Generations
  • First - More of body-shopping , providing manpower for client's tasks.
  • Second - is the BPO we know, took client's tasks(mainly contact centre ones) to offshore cheaper locations.
  • Third - Transaction processing for the client. The pricing model changed from input based to transaction based. Eg: Xansa used to handle metered billing exceptions for U.K based utility company ThamesWater.
  • Fourth - The current wave, involves analyzing trends across the transactions and make/suggest process changes for business benefits. The transition is happening from non-core transactions into more core ones, thus partnering with the client for their business strategies. The job is no more easy. And we have CA's , MBA's , statistician joining in for understanding and modeling these trends. There is no doubt that BPO is a people-intensive business but the model has evolved out of growing linearly by increasing the headcounts to providing real knowledge and skill based solutions.
  • Fifth - coming up , involves leveraging the social media & cloud computing as service differentiators with focus on location based analytics and specific catering to SME's scaling their operations.
So if you are getting into just any job in BPO industry , make sure of the above changing trends (yes , the line is getting blur between the core IT jobs and BPO/KPO jobs) and select your area of interests accordingly and work accordingly to gain the relevant expertise. Its a huge market; as per NASSCOM , the Indian IT-BPO revenues is projected to touch US$ 225 billion by 2020. The BPO industry alone is expected to touch US$ 14.1 billion in 2010-11 , still a double digit growth of 14 percent !
Back to Top