Tokopedia - My first time outside India

It was August of 2017 when I received a text from Ankur Malhotra (ex AVP Paytm and then VP at Tokopedia) to join Tokopedia in Indonesia as a Lead Product Manager for Marketing Products. I was not sure, being an Indian I never had exposure to Tokopedia, except an article in Techcrunch, where they had raised a billion dollar just that week.
‍

I was reluctant to say yes, but I knew Ankur was a smart guy and I had wanted to work with him since Paytm. We discussed the roadmap and I was excited about the products, none of them which I had worked on in past. It will be a new learning experience and a chance to move away from the payment industry and consumer products. Also, being a traveler, I love meeting new people and seeing a new culture. I said yes to Ankur next day and joined Tokopedia on 25th September 2017.

“MAKING A BIG LIFE CHANGE IS PRETTY SCARY. But, know what’s even scarier? REGRET.”

I still remember landing in Jakarta, full of excitement and nervousness, due to the language barrier, couldn’t speak to the driver. I was just looking outside the window, driving to my hotel, I saw a hoarding of Tokopedia, an instant smile and goosebumps ran across my body. I think saying yes to Tokopedia has been the best decision of my life so far.

**Few things about Indonesia and especially Jakarta before I dive into Tokopedia. It’s been more than 18 months here now and I have learned so much here as a person. Coming from India, where people have the “Me First” mentality, Jakarta was different, no honking on the roads, people giving way to others (Though Jakarta traffic will kill you!!). People are so respectful to each other, irrespective of your social status (I have seen millionaires being polite to Gojek drivers). People are so humble and I learned to say Thank You and Sorry here. I was living alone here and thus learned how to be independent. I was able to manage everything on my own, so not just professionally, but personally, I became more mature. I made some really great friends and connections here, which I will remember lifelong.**

Now back to Tokopedia, I remember the first day meeting my colleagues, I was nervous, like being introduced to a new class, only that everyone looked different and talked differently. My colleagues Igor and Holy did my onboarding and presented me with few challenges faced by them and products they need to run marketing effectively. These were –

Referral Engine
Product Feed
Data Integration
Chatbot
Affiliate System
Deep linking

I will discuss these in details, but I was able to fix these in the next 6 months. I learned about the tribe system here, which was borrowed from Spotify. It took me some time to get used to the system, know people and get used to the culture. I realized one thing quickly, me being hired from India put a lot of performance pressure. Other Nakamas (word used for Tokopedia employees) looked up to me as if I was a lot of qualified and knew everything. Partly, this was because I was from Paytm, which was very successful in India by now and Tokopedia was looking up to it for few inspirations. This was a good thing since it forced me to go an extra mile to learn things, I took few online marketing courses on Udemy, Linkedin and Skillshare, to learn FB, Google, SEO, Affiliate and email marketing.

So now starting about the work, I have undertook a lot of initiatives at Tokopedia, which mainly aims in acquiring new customers, engaging current customers, helping marketing team in their work and lastly few data products owing to my interest in data science.

Deeplinking – The first project which came to me was solve deeplinking in app, we were driving huge traffic to mobile site at that time, which had very low CVR, the links used for app were not working correctly. We found solution in Branch.io , a deeplinking solution, which we implemented for our sharing feature in app, we also started using branch deeplinks in our product feed, in our emails, and for any other link we shared on social media. We were also running Web to App journey and were able to achieve following results –

3X increase in app downloads using Journeys
30% lift in revenue as a result of using Branch’s Universal Email
28% growth in total MAUs
1.5X increase in iOS and Android transaction rates through deep linking
30% reduction in misattribution of transactions across web and app

More can be read here.

Referral System – The next project was Tokopedia Referral system. The system was made to reduce our customer acquisition cost and drive app install. I will not be quoting the numbers here, but this channel was one of the biggest contributors to our new users. It also helped reduce CPI for our app and the retention of people from referral was better than customers acquired from other channels.

LISA – Other problem which marketing team was facing was on product feed. Product feeds are a catalog of products you have, which you upload on Google, FB or other partner sites, which then run ads based on that. This feed contains columns like product id, product name, product image, price, landing page etc. Earlier this feed was requested manually, to an engineer, who will then query the DB and generate feed and share it to the marketing team. This whole process as you can guess was inefficient and took sometimes 2-4 days. Also, since the price of products might change, and products could go out of stock, we needed to change this system and automate the whole process. I worked with Suman, a very good developer from Tokopedia India office to automate this whole process. We came up with a tool we called LISA, where we can put filters for feed, select which format is needed (FB, Google, General, Microsoft etc.), then type (CSV, TSV, XML). Also, you could set time, at which feed will be updated and made sure the feed is updated on the same link, thus no need to change link every time. The partner could just refresh the feed once a day, and you have updated products. This reduced feed generation time from manual 2 days to automated 2 hours or less. There were also, some more advantages of this system over the previous system, which were more on detecting nudity and removing products with certain keywords, categories, and shops.

‍

Tokopedia Go – Google was planning to launch Android Go platform and we were selected as few beta developers. We optimized our App to run on Android Go platform, the focus was mainly on reducing app size, reducing background process to reduce RAM consumption and also battery optimization. The code was refactored and we were one of the first Indonesian app available on Android Go platform.

PWA and AMP – Next in line was another project in collaboration with Google. Indonesia had a lot of mobile users, who were visiting on mobile site, but were not ready to download the app. We wanted to improve experience for these users and for same we started preparing for PWA and AMP pages for Tokopedia. The case study can be read here. Main out puts were –

5 times higher conversion rate
60% less bounce rate
65% faster page load speed.

It was hardwork of our Lite tech team, who was very serious with this project and Dendi our EM for lite was personally paying attention to this project, so kudos to them.

Personalization and Recommendation – I had been developing an interest in data science for a year now, and had done few courses on Coursera (Andrew Ng), DataCamp, Udemy and Udacity on Machine Learning and Deep Learning. I was scrolling through Tokopedia website and startled to see a recommendation for women products, now having knowledge of recommendation systems, this was not acceptable, and I discussed with Manoj Awasthi, our VP of engineering about the problem. I started working with data scientists Laurent and Nishat, to improve the system. In this journey, we faced and solved problems like user clustering, data availability, product clustering and so on. We developed a few models, each serving a different purpose.

Prediction – Given a category, which users will be buying in it in the next 7 days, based on their behavior on Tokopedia in the last 30 days. This problem was interesting since it was also being worked in parallel by Google team, and we were trying to optimize our ad spends and user targeting.
Recommendation – Based on user session, we are making a recommendation of products to users. This approach is based on the assumption, that in a given session, the user views similar products or products that are bought together. So a user who is buying iPhone will look at different smartphones or will look, buy or search covers for iPhone.
Target time – We also worked on a model to identify the best time to target the user for marketing (push and email), this as of this writing is work in progress.

The goal of all these models was to improve user experience on Tokopedia, increase CVR, optimize marketing and make Tokopedia more personalized to users. It was a very good learning experience personally because the conversations with Manoj, Youssef (VP of Data), Laurent, Nishat, and other data scientists taught me a lot. I was always curious and used to google and read the papers on all the new terms which I heard.

Data Hub – Being one of the biggest e-commerce in Indonesia, we were generating a lot of data from different platforms (Web, Lite, Android, iOS, Server) and sending this data to various destinations (CRM tool, Appsflyer, Branch etc.). This consumed a lot of developer time and also was a blocker for innovation, for reasons that can’t be mentioned here. So I proposed making a data hub, which will be a single source of data, all the sources inject data there, and we can manipulate data before sending it to any 3rd party via API. The data engineering team Sonny, Yunita, Sulung, and Dicky worked with full force to get this system up and running. Thanks to push from Youssef and his buy in for this project, this became one of the companies key initiative for 2019.

Sitemap – SEO is one of the key traffic sources for any website now and in e-commerce like Tokopedia, which is having different verticles like a marketplace, flights, digital recharges, tickets and deals. It becomes very important to help Google index those pages, I worked with the team to enable automate our sitemap (adding any new product automatically to robot.txt) and thus increasing our visibility on SEO.

Android App – While at Tokopedia, I was also, working on optimizing our app and making it fully compatible with Google guidelines. Thanks to our tech team lead by Rico and Vishal, we were able to get our app two accolades –

Google Play User Choice Award 2018
Google featured App for June 2018

Tokopedia Challenge – Indonesia is a very social media driven market, you have to be here to see how much is Instagram instilled in daily life of people. Ankur came up with this idea and I helped implement it, he wanted to increase organic virality of Tokopedia. He suggested to make a platform, where social media team will run challenges(Like unboxing video, couple selfie, get fit etc.) , users will participate in those challenges and in order to win the challenge, they will have to share the links to these challenges on social media. More the interaction (likes, comments and share), more are your chances to win. A score was generated for these interactions, and winner was decided based on highest score. It was a very interesting and fun project, Heena from my team was working on it. More about my team in coming paragraph.

Marketing Automation Tribe – One of the best experience of Tokopedia was to build a team from scratch. With growing number of projects, it was becoming more of necessity to have a full dedicated team, for which Ankur helped me to set up a complete dedicated team. The team included backend engineers, Android engineers, iOS engineers, frontend engineers, first time I lead 3 dedicated Product people directly, and 2 data analysts. As of writing this article, we are working on a lot of cool projects ranging from Augmented Reality, Marketing Automation Platform, Tokopedia Blog etc.

‍

Working at Tokopedia has been one of the best experience so far. I made so many new connections. I grew personally and professionally. I spoke at 4 different conferences internationally. I learned about marketing and data products. I learned how to manage people. Working with different culture and language was not easy, but the people of Indonesia have made it a breeze so far.

‍