A Series: The perfect groom for the run-away bride

MUMBAI, INDIA: Three guesses. That's all you get.

What is the common denominator between the fierce pair of peppy writer Richard Castle and prissy detective Kate Beckett and - the adorable Rishta duo of impish Rohan Mehra and sweetilicious Isha Mirchandani?

If you are googling their names now, we will save our what-hole-you-have-been-burrowed-in kind of disappointed look for later. Go back to your books or space station or whatever it is that makes you insult the holy idiot box.

For others, have you uttered the word ‘mystery-solvers' already? Not wrong, but no, that's not the code we put on the hook.

Did you just say ‘the magic of opposites'? Impressed overwhelmingly, but sorry, you have to try again.

Tick tock, tick tock.

Wait, sounds like we heard a quiet sigh suffixed with ‘why don't they get together, they are so perfect as a couple'. Aww, here's a tissue box. We totally empathise and will join you in that heart-rending dissection, but for now -alas, you have exhausted all your chances.

Ok, so here's the answer. Both the pairs are armed with a lot of processes, savvy-teams and sci-tech-paraphernalia and yet what plays the clincher in solving every case, whether it's a whodunit murder or a relationship-fog, is one simple old-fashioned magic: human insight.

Insight, intuition, guts, sixth or seventh sense, deep understanding, a grip of the just the right pulse, an incision into the exact vein- no matter what you call it, thanks to the powers of Universe aloft - the magic still works.

Nothing can replace it, and yet technology can complement it beautifully. Speaking to Gourav R, COO, Shaadi.com makes you shake some dust on this long-lasting theorem as he gives you that extra minute to absorb his wise lines " The truth is out there, you just have to ask for it."

Chances are that by now, this sounds like a nail-biting case already. So if you are game for a good episode of how can analytics win or fail as the engagement ring for users in matrimony industry, let's switch it on. And no, it won't take 60 minutes. Join us.

 

For someone who tries to match technology, operations and business model seamlessly, what has been your reaction to the much-touted word ‘Big Data'? How can anyone make sense of it without being overwhelmed or left at the station?

While we may build hypothesis and use intuition where it is well-slotted, a good solid framework is still relevant. We anticipate that customers, users and other constituents in this industry can sometimes show completely different behaviours. That, however, is not a problem but an invaluable asset. Algorithms and frameworks have been around. Analytics helps in building continuous intelligence though. It can take existing models to new levels.

How is that possible without letting technology commoditise your propositions? There are so many portals and dating or networking sites popping up as we speak.

Technology is a commodity. But IP is the differentiator. It is still a game of who uses it smartly and strongly vis a vis competitors. Being able to generate an insight is a real competitive advantage. It extends the distance between a strong player and its rivals. The How-to and the ‘human' part still matters tremendously and one has to understand that there is a lot of difference between data and insight. This is a different line of thinking compared to using just some black box.

Give us a peek into your view and use of analytics so far?

For analytics we use a variety of tools and technologies, each one selected for the specific nature of problem we are solving. The analytics tools we work off include Excel, Business Objects, Pentaho, SQL, Red Shift, Hive/Hadoop, Google Analytics, SAS, Kissmetrics & R along with proprietary analytics tools and frameworks developed in house to support or Insight needs. We keep moving to newer features and tools as required and our entire stack matters. In the last five years the landscape of analytics has changed quite considerably and credibly. The advent of Big Data is an important factor but what is remarkable is that now we are able to see many stable technologies and solutions. The accessibility has also improved a lot. They are no more locked up in data warehouses or for use of data scientists only. The entire data mining landscape is all about the right tools.

But the interesting thing, as you said in an earlier conversation is that - no matter how great a bird's eye view IT allows, you still have to pick up the phone. Explain that contradictory streak.

We still put a lot of emphasis on meeting our customers, and keep spreading our centres and face-time for that good-old customer-connect. There is a vast network today of our presence across the country and that validates our focus. I feel that beyond a point, some old stuff still works.

Intrigued. Please elaborate.

See it works like this. All these algorithms and tools for matchmaking are still about that discovery process - the one person someone is looking for. We cannot sideline intuitive and experience-based or behavior-pattern-founded lessons no matter how automated everything can become. Our job is to help that person make this discovery. Our expertise is not just out of some model or process or giving the best way to drill out a data, but about making the journey to that ‘one person' easy, smooth and not-too-overwhelming. I still believe that the truth is already out there, and all one has to do is ask the right questions.

 

You remind of Mr. Miyagi. Is there still some wax-on, wax-off time left for technology before it can master what you are expecting? What move is missing?

We are still looking to marry deep-analytics with real-time customer response. It has been great to use technology in a lot of offline work, but in terms of real-online spontaneity this is still lightweight and a bit diagnostic, especially when we expect the desired maturity. To do what we do offline in a real-that-very-second kind of way will improve everything exponentially.

Any update on the progress made in areas like collaboration or blacklists towards the goal of fraud mitigation?

It's a continuous process and we take it very sincerely and seriously. That is a journey and not a destination as I see it. We use large set of diagnostics for now. We can use patterns and gains of IP, codification, training etc to a great level to protect the interests and feelings of our members. The challenge of pattern-matching is crucial in stopping people from repeat misdemeanor. Thinking of that, perhaps, industry collaboration is not too far-off either.

In a nutshell then, how would you tag technology in this space?

A big learning for us is that if we do not define the problem before using any solution, no matter how sophisticated, it just won't be good enough. Data is massive and all around but it's the how of wielding that data. It's a rat hole of sorts and it can still take you to places, but the point as always is - you have to know the ‘where'. Common sense is still not out of fashion here.