My Keynote Presentation at Data Science Conclave 2017 in Chennai

I am sharing my my keynote presentation at Data Science Conclave 2017 in Chennai. Thanks Rajesh for the invite.

1. Major improvements in accuracy in speech recognition and image recognition opens up a new field in human computer interaction. With computers able to correctly interpret almost all interactions without direct contact with keyboard or mouse, a major data source has opened up for Data Scientists to explore.
2. A system which is 80% accurate may not usable, however, when accuracy crosses 95%, there is a major turnaround in large scale adoption.
3. Self driving cars will lead to major leaps in technologies for object recognition -> not just previously known objects, also to anticipate and correctly handle unexpected objects.
4. In my view, there are four key dimensions of Data science, these are Data, Domain Expertise, Machine learning algorithms and Technology of Deployment. Value creation is possible across all the dimensions of Data Science. Better quality data, higher volume of relevant and contextual data can create value, and domain expertise remains critical in making successful deployments of data science projects. Our focus on machine learning algorithms is important, however, value creation happens across all the four dimensions.
5. We have seen a 5X increase in jobs which require machine learning and neural networks expertise.

Data Science is now mainstream and it is important for every organization to invest in Data Science and benefit from it.

https://www.slideshare.net/vjain99/data-science-conclave-keynote-presentation

Customer Insight - Survey, Data and Interviews

Customer insight is the only sustainable basis for building a business. And in my humble view, there are three ways of gathering customer insight, 1. Conduct a survey, 2. Analyze behavioural data, 3. Conduct in depth customer interviews.

There is a lot of behavioral data on the internet.  From browsing history to clicks and transactions. However, behavioural data only says "what" is the customer doing and when. "Why" is missing !!

Customer interviews can give insight about all that a customer cares about. The only challenge is that it covers only one customer at a time. Often, you may not have a large pool of customers to do indepth interviews, a lot of time may elapse between successive interviews or different customer context may make it hard to correlate insights across them. That's where surveys come into play.

Surveys enable quantitative measure of customer's opinion. If you have access to a large pool of customers, surveys are low cost and quick. Without surveys extrapolation of 1/few customer's opinion as the fact, is fraught with risks.

And without depth interviews, survey insight is shallow.

And without correlation of behavioural data, surveys and depth interviews are just opinions, which may not show up in reality.

If you are wondering on how to move from "I feel" to "I Know", use the three methods of gathering customer insight. You will know when you can say "I Know".

Digital Marketing and Relevance

For any brand, digital marketing is now mainstream. Ignoring digital is like ignoring reality and for a marketer, almost a professional suicide. When brick and mortar businesses like hotels and restaurants have to rely on digital marketing for business, no one can really remain untouched.  Any digital marketing initiative must examine relevance of message to the recipient, one it is possible and two, your business will cease to exist without it.
There are learnings we can derive from already established digital businesses. Take ecommerce companies for example, mailers drive traffic for major offers and seasonal discounts.
1. Capturing sufficient information to personalize the message - Relevance has multiple connotations, every person has unique set of requirements and aspirations. Without capturing this information relevance cannot be achieved.
2. Current and correct information - Capturing detailed information that is correct and updated is also an important challenge. Customers provide their location and interest in products and categories through their browsing or purchase behavior. It is obviously unrealistic to expect every declaration of interest to match with displayed behavior. Hence a strong need to constantly update the customer attributes.
3. Classification - Another tool for matching that is used very often is classification. For example, an ecommerce shopper expresses interest in bed sheets, now does that mean we can send offers on Garments or home furnishings. While classification helps find commonality between both parties, classification errors add to the complexity.
4. Error tolerant matching engine -
A vertical specific matching engine can build on the domain expertise present in the organization. To make it more robust for errors in data capture and classifications, matching engine may use behavioral information or the domain expertise.
Improving relevance finally becomes a function of improving all of the above. Do share your thoughts. Wish you success in your digital venture.

Story of a Product Revamp - Naukri Resdex Emails

If you are looking at apply for a Product Manager position at InfoEdge and will like to know the kind of work this team does, take a look at the list of changes we made over the last two years to an established product.

Resdex is the Naukri Resume Database product, which enables Recruiters to identify relevant jobseekers for a specific opportunity and contact them. Recruiters can either call up the jobseekers or send job opportunities as targeted mails. Resdex mails is a popular mechanism of contacting jobseekers and has undergone a major revamp in last couple of years.

1. Addressing Relevance for Jobseekers - Resdex allows recruiters to search jobseekers based on several criteria like key skills, designation, company names, salary, experience, location, education and other criteria as well.  However, Naukri was getting several complaints about irrelevant mails being sent to jobseekers.  Typical complaint being "I have done an MBA from premium college and I am getting call center jobs".  To address this challenge - Naukri introduced mandatory filters for Salary and Experience, however, the choice of the values was left to the recruiters discretion.  We saw a major fall in the volume of email sent (and loss of business) as recruiters started entering the expected salary and experience consciously.  This reduced the irrelevant emails sent to jobseekers who were never the targeted audience in any case.  And since jobseekers saw more relevant mails in their mailbox, the overall response to Resdex mails improved significantly.

2. Apply and Reply from the mail - Naukri has now enabled Apply and Reply from the mail.  Jobseeker can login to Naukri account and apply to the job sent, Or compose a response and attach a resume as part of reply.  The recruiter receives the jobseekers Naukri Profile snapshot along with the resume, the profile snapshot is a standardised template that enables a quick scan of applies. A large number of jobseekers are now checking their emails on the mobile phone and where they may not have access to resume, it gives them an easy way of applying through Naukri and using the uploaded resume to quickly respond to the recruiter message. 

3. Recruiters can view list of Contacted Candidates and Applies by the Job Sent - Contacted Candidates information is now displayed subject line wise (which in most cases represents the Job Title), recruiters can now see which candidate was contacted for which opening. The number of candidates contacted is indicated against each Resdex email subject and link against each subject line can be clicked to view applies received for that email in EApps. Since jobseekers now apply through Naukri, list of applicants is also available to recruiters.

4. Send a Job as Email - After posting a job on Naukri, you can search the most relevant candidates in Resdex and send them the job. Send a Job as Email thus combines the power of Job Posting and Resdex, and helps in closing the positions at the earliest.

5. Automatic Shortlising of Candidates - Searching and shortlisting of relevant candidates from Resdex may require significant amount of time and effort. Now, Naukri can also help in identifying the matching jobseekers. To begin with, this functionality is available if you have posted a job where annual salary mentioned is more than Rs 15 lakhs per annum. The candidates are selected using the Naukri iMatch technology which also powers Naukri Job Alerts. The matched candidates are available in a folder in Resdex which carries with the same name as the job title for convenient access. If you want, you can further shortlist candidates from this set or send the job as email with a single click.

Who made this revamp possible - of course, the product managers, along with the User Experience Design, the Technical Team and the Analytics team.  Kudos to Abhijeet Anand and Praveen Chandran, the two product managers who have led this revamp.

Naukri Referral Hiring Product

The new Naukri Referral Hiring Product enables organization to manage the Employee Referral program with considerable ease.  Here is why I think this is an amazing product -

1. Easy sharing of the referred job by employees on their social networks -  Employees can share the jobs on their social networks with a single click.

2. Tracking of applies - Applies can be tracked even if a friend of friend of the employee applies.  That is, if a job is shared by an employee, a friend likes the job and friend's friend applies to the job, employee will get credit of the apply.

3. Management of contact lists - You can manage the list of employees to send the referral mail to. For example, if it is the sales manager job, it can shared with employees in sales team, who are more likely to have sales managers as friends.

For more details, check out the Naukri blog - http://recruiterzone.naukri.com/?p=3173

Do give it a try and share your feedback.

Story of Naukri Job Alerts

Naukri.com is the market leader among with career sites in India, with market share currently at 63%. Naukri.com has 30 million+ registered profiles and a large part of these registered members receive a job alert every alternate day or on a weekly basis. Job alerts only contain freshly posted jobs on Naukri.com in last two/three days. It is probably the main reason why Naukri Job Alerts have one of the highest open and click through rates. Yet, jobseekers complained of relevance of jobs sent. That was identified as one of the important problems to address in early 2010.

The process of improving the job alerts was an incremental one. We built the logic step-by-step and with every incremental step, our understanding of the relevance problem improved.

1. Discovery of “Role” – I tend to believe one major variable than we discovered was “Role”. A deep dive in the behavioral data showed several interesting patterns. Jobseekers were not sticking to their Functional Areas (departments) and were applying across Functional Areas.

a. Pattern of apply clearly indicated that Role was more important than functional area.

b. We had roles which were very similar present in multiple Functional Areas, for example, sales role existed in Industry specific functional areas. GM Accounts existed in Accounts Functional Areas as well as the Top Management Functional Area. Also, several functional areas were close to each other, for example, Accounts and Banking.

2. Limitation of Keyword search – Key skills entered by jobseekers represents what they consider as important. Logically, a search on jobs should use the key skills entered by the jobseeker. However, some of the jobseekers had not entered their key skills. A large gap existed in the key skills entered and their skills as apparent from the CV. We needed a robust mechanism which did not fail because of the data inadequacy.

3. Handling of Categorical Variables – When we compare two jobs and their relevance to the jobseeker, attributes like “role” were important. The key challenge was to translate this into a distance function that can be used in predicting relevance for the jobseeker. Similarly, attributes like Industry, Location required identification of a good distance function.

4. Jobseeker Resume – A match between a jobseeker’s expertise and the requirements from the recruiter is essentially a match between the CV/resume and the Job Description. Of course, there are challenges – if a CV is old or a job description is incomplete, this may not work very well. Yet, we needed a mechanism for matching the candidate CV and the job description.

5. Apply Behavior - It is very much possible that apply behavior of a jobseeker will deviate from the CV/resume.  Apply behavior can provide insight into asiprations of the jobseekers as well as help identify classification errors. Incorporating apply behavior in identifying matching jobs for jobseekers is another significant challenge.

Naukri Analytics team identified the above challenges and incrementally solved them in association with the product team and the technology team. And of course, we noticed a major improvement in relevance feedback from jobseekers.

We are not done on solving this technical challenge. Analytics team is looking to hire smart Data Scientists to join its rank and work on solving these – if you are interested, please click to here to apply.

Information and Communication Technology (ICT) Revolution has Touched Everyone

Information and Communication Technology (ICT) Revolution has delivered on its promise.  While our expectations on what it can do are still very high, I believe we have achieved a lot already. Communication revolution has given IT a platform to make life easier for almost every Indian who can be reached via road, rail, on foot or by the mobile network.  I share some of the use cases and associated stories with you, as in my opinion, they provide a vivid picture of how IT and Communication has changed our lives for the better.

1.       Travel and Ticketing - Most visible benefit of computerization that common man saw in late eighties was “Computerized Railway Ticket”.  With charts displayed prior to boarding the train and real time status of railway reservation, in one stroke, it took away the opportunity of rent seeking behavior. Millions of Indians, who travel daily on Railways, not only saw the benefits, but also became the champions of IT in India. That, to my mind, is the single most driver of IT adoption in India. 

2.       Travel and Ticketing - Internet ticket by Railways is also nothing short of a revolution. A local Internet Café accepts payment in cash and delivers me a ticket without the need to approach the local railway station. Savings in cost of travel to the station and back, no time taken to travel and easy access to railway information, all benefits have been delivered with the convergence of IT and communication.

3.       Real time Communication - 20 years ago, an Aunt of mine was coming to visit us. She lost her way and we learned that she had left her residence a long time ago and by all means, should have reached us by that time. Yet, there was no sign of her.  We imagined that she had probably lost her way.  We went in several directions to look for her. An hour long search and we could locate her, safe yet tired.  Just imagine the scenario today, she carries a mobile phone. We can reach her anytime and she will probably never get lost.

4.       Location Services - A similar incident like the one before. A few years ago, while traveling by train, I woke up with a start. My train was expected to reach by destination at 4am in the morning. It was already 3.40 am and I did not know how much time we had. I switched on my Blackberry and checked my location on Google Maps. We were just 2 Km away from the destination.  The train was running ahead of time. I knew we had to rush and get ready before the station arrived. Google map and easy access to mobile network on a train route helped me find the exact location and prepare me.

5.       New Customers for Small Businesses - Until a few years, for any problem with electric fittings or plumbing requirement, I will visit an Electrical or Sanitary/Hardware shop. I will wait for an electrician or the plumber and leave a message for him to visit us.  A number of follow up visits later he will arrive and do the necessary job. Today, every electrician, plumber or even a domestic maid, carries a mobile phone. They are reachable immediately and confirm their availability.  Everyone has benefited in the process. Of course, except the Electrical or Sanitary/Hardware shop owner, who has lost the cut he used to get.

6.       Mobile Payments - A construction worker was due to receive his wages from the contractor. However, he asked the contractor to keep the wages in his safe custody as he was afraid of losing the money since he had no regular place to stay and risk of thefts. A month later, when he wanted to go back to his native place, contractor made him run around for getting the accumulated wages. With no place for safe keeping, he was at the contractor’s mercy. That has suddenly changed with Money Transfer on the mobile phone. He can now transfer money back home without waiting for his next visit.

I am sure all of us have experienced something similar in our lives. ICT has touched everyone. With new innovations, our expectations are also soaring high.

·         Touch and speech are two new interaction elements which are making IT more accessible to every individual.  Touch phone is more intuitive and so, is speech. Siri is a great beginning in this direction. Speech recognition as a technology holds great promise and with a commercial value attached to quality, I am confident it will soon improve to serve the masses.

·         Another promising initiative is Aadhar. That shall help the common man in opening Bank accounts by providing an identity proof required to access services.  Will banking correspondents make up for absence of banking branches? That remains to be seen.

·         Will remote healthcare delivery graduate from being pilots/experiments and scale up to District or State level delivery models? Healthcare workers cannot replace the expertise of a doctor, yet they can at least provide timely help for the patients.

ICT is a force multiplier, yet it is not sufficient to improve the living standards of the common man.  Without roads, it is hard to make services accessible in the hinterlands. Without drinking water and access to healthcare, progress will remain limited. Physical infrastructure is a must for progress.