Thursday, August 6, 2015

Image Search, Ecommerce and Bloggers ...

In Part 1 of the series, I posted about what went behind the scenes to attend the #FlipkartImageSearch experience meet, powered by Indiblogger team. In Part 2, I would like to discuss further about the Image Search feature recently launched by Flipkart and how entire eCommerce eco-system can benefit once this feature evolves.


Image Credits : Indiblogger

Every technology advancement undergoes immense scrutiny and likes of Google, Facebook and Twitter ( hmmm, to certain extent, from those early days) have served us pretty well . Image search is not new and is in use for long time. However, Point,Shoot and Buy of Flipkart is repackaged to suit its core business - selling products. Its immense matter of pride that its Made in India led by Puneet Soni, Chief Product Officer . If you are hiking in hinterlands of rural India, find an awesome dress matching your taste, you take a snap, search it and buy. That's simple, right ? Hell yeah...
That would be the ideal situation which Flipkart's team would be hoping for , and infact,we the end users would not just expect, but also demand it.

Yesterday, I reached out to my wife over chai conversation and she was more than happy to be my soundboard . Few of the points below came up and I thought to put it down too.

1. Novelty - Amazing feature if marketed well. By stopping millenials to click selfies and turning them to Point,Shoot and Buy, entire smartphone userbase with embrace it. After all, we are one big market for the eCommerce players.

2. Relevancy -  Books, gadgets might give better results when their images are searched but its usage will be primary clothing and fashion accessories I believe. Why would I click a book's picture when I can type the title of the book or phone in text search box ? Though the accuracy is better in former, I would like Flipkart to improve image search overall. Getting this right will do so much good to their business and getting it wrong might be similar to the scales of #bigbillionday sale goofup on long-term basis.  Imagine Snapdeal gets it right, your endusers ,who have forgotten #bbd sale,  may not forget this ever for getting it wrong.

3. Specificity vs Variety  : When we were engaged in team activity, Punit came to our team(Vyapam) to see how we were doing. I asked him that if I click a photo of my bright-colored striped shirt and fastrack watch, which product will come up in search results. He straight-away said that it would be the shirt because of its size and color contrast. Theoretically, fine ! But now, when it sunk in, it's really weird because when you boast of 25 million products, why do you limit your search to something which looks good in the picture? Instead why don't your results display both the products - shirt and watch - and motivate the buyer to buy both. This conversation was totally off the record and we both might have missed the nitty-gritty.  So guys, don't take this to your heart and consider it as viewpoints of two individuals.

What next ? Enhancement requests
4 . Obviously, Voice Search or BOLO,DEKHO aur KHARIDO or, ASK LOOK and BUY ! :)

5. Product availability notification -I heard from one of the flipkart guys making a statement that they retain images for short while when my fellow indiblogger questioned about privacy concerns. Instead, retain it for extended time -maybe 3 months ,  get to know the customer intent better by her images uploaded in the app and target for future sales - maybe informing the buyer when you find exact match, if it was not the case at first place. Or creating a personalised catalog of images in a mailer, instead of those promotional mailers which we move it to Promotions tab of Gmail and treat them as ugly sibling of Spam folder.

I am sure your excellent team of product managers and engineers have given out these suggestions and many/all the above are in pipeline. Maybe in your next few sprints, you will deliver them too.  After all, we have come a long way in searching for books in crammed Avenue Road of Bengaluru to getting the book delivered at our doorstep. This post was only intended to stop naysayers about slamming the product and adopting a more patient approach towards e-tailers to deliver magic we order .

Endnotes : As a database engineer managing large-scale enterprise infrastructure, it was quite pleasing to hear from Punit that their team has built their infrastructure to such an extent in last few months that #bigbillionday goof-up is a non-event if it was held today. Wish you good luck guys for the #TheFreedomSale !

No comments: