With over 20 years of experience in driving analytics and customer insights into strategic planning and decision-making in a wide range of organisations, Ashish Singru currently leads eBay Bangalore as a Senior Director & Head of their Global Business Analytics Centre. In his current role, Singru has set up capabilities to help solve eBay’s hardest business problems as commerce and retail worlds converge between online and offline.
The key business domains for which he has driven strategic planning so far includes areas like customer engagement (CRM), sales & marketing effectiveness, media planning & buying as well as product design & development.
He has partnered closely with senior corporate leaders in major retailers, telecom, automotive, consumer good & technology companies to create solutions to address short-term and long-term challenges and opportunities for the business using a data-driven approach to analyse customers, products and markets.
Prior to eBay, he has worked with the likes of Microsoft, SABMiller, Young & Rubicam Group and Gallup Organisation. He has a B.Tech. degree from IIT-Kanpur, MS (Rutgers) + MBA (Iowa).
In an exclusive catch-up with Analytics India Magazine (AIM), Ashish Singru shared his insights on the role he plays at eBay, adoption of analytics in E-commerce, future plans for analytics, challenges and much more.
AIM: Since your stint at eBay as the head for the global analytics center in Bangalore, what major developments has the company witnessed in analytics domain? Would you like to explain in details about various analytics development?
Ashish Singru: I would like to call out 5 areas where in my opinion lot of advances have happened in the past 5 years during my time at eBay. Firstly, our Experimentation methodology and related technology capabilities have advanced quite a bit. For any internet company, having a robust Experimentation infrastructure is critical to understand how to improve customer experience and also refine marketing strategies. Strong analytics is helping eBay evaluate outcomes of A/B tests accurately and take good decisions. Applying technology to the Experimentation process helps analysts focus on generating insights and less on data issues.
Second area is applying analytics to User Experience optimisation. Essentially, it means using the right metrics to continuously track if our website features are working as intended, and quickly identifying issues and resolving them. E-commerce websites and business models have lot of interconnected parts – you touch the business in one area, it has ripple impact in another – and we have developed a rigorous monitoring process that helps product managers and engineers address issues and helps business teams understand impact of changes to the business outcomes like GMV and User Growth.
The third area where we have seen lot of progress is financial analytics. At E-commerce and internet firms, there is rich data on transaction volume and revenues on virtually a daily basis, and lot of opportunity to build sophisticated models and analysis to both explain and forecast business performance. Analytics is a very strong partner to the Finance function in this regard and helps them mitigate risks and amplify opportunities to use financial levers to move the business. We also have a rigorous approach to understand impact of external factors – such a Weather changes or Extreme Weather events, Macroeconomic events (e.g. Brexit) and day to day currency fluctuations.
The fourth area is Trust & Risk. E-commerce and Marketplace models must have strong policies and user experience features to ensure that both buyers and sellers can transact with confidence. Analytics helps evaluate/simulate various policy options before they are launched on the site to determine the optimal one – which balance multiple outcomes. Also, when the policies are embedded in the site in form of web features, analytics helps understand how user behavior is influenced by the changes. Once changes are implemented, analytics also helps monitor and evaluate impact and ROI of the policy and product changes.
The last area where I have seen lot of development is self-serve analytics tools that business can use without needing the analytics department to help them day to day. This is also an area where we have seen very impressive innovation by analysts, who start with an idea on an elementary Excel spreadsheet and it evolves into a production level system at enterprise level available to hundreds of users in the company – whether for inventory management, or pricing or advertising. Exciting stuff!
AIM: Would you like to highlight a specific use case in analytics that has brought significant value to your company?
Ashish Singru: One example, which is related to the retail nature of E-commerce business is Deals. When we offer a Deal to a consumer, it has to come from a seller on the platform. While sellers use their own budget to provide deals, e-commerce companies like eBay also help subsidise the cost of a deal for the seller. Analytics is very helpful in determining how much should eBay provide in terms of deal subsidy to have optimal conversion on the deals. It is a win-win for sellers and eBay to have that insight to ensure that deal is priced attractively for buyers and the seller and eBay get sufficient ROI from the offering – helping everyone win!
AIM: How well is India faring in terms of analytics adoption in e commerce? What is your take on adoption of other technologies such as artificial intelligence and machine learning in E-commerce?
Ashish Singru: E-commerce is one of the sectors where analytics adoption is high among Indian companies. Hopefully, Indian E-commerce companies will help lead the way on creating data-driven cultures in Indian businesses. Kind of similar to how Financial services companies did that couple of decades ago in US. AI and Machine Learning will be critical to offer differentiated customer experience on the website, as well as in helping these companies efficiently manage Marketing, Trust/Risk and Inventory. In Indian e-commerce companies, the focus has been a bit more on Operational Analytics – that helps answer questions day to day to run the business – but clearly there is a trend to hire the right kind of talent for AI/ML and using it. Using the “horse before the cart” analogy, what is critical is to make sure we have well defined problems that are best solved using AI/ML and then go solve them using these advanced techniques.
AIM: Is eBay upping its game in terms of adopting these technologies?
Ashish Singru: Yes, eBay is investing a lot in these technologies, and our engineering centers around the world are building up their skillsets in these areas, to solve the problems in domains I mentioned earlier.
AIM: What is the hierarchical alignment (both depth and breadth), of your analytics group?
Ashish Singru: While a lot of analysts are aligned with the Finance function at eBay, we also have analytics teams embedded in product, marketing and risk organisations. Every day more and more functions in the company are embracing analytics, so the breadth is growing as well.
AIM: What is your roadmap/plans for analytics at your company in the future?
Ashish Singru: One area where we focus a lot is to become better at storytelling using data, so that one can influence the business faster and drive better actions from the insights. Secondly, increase the use of data science technique to solve hard problems and to also build tools or products that can be embedded into our site and business processes and can work in an automated manner. Third, grow business acumen of our analysts so that problem solving using analytics can be speeded up and things do not get lost in translation between data scientists and business teams. Increased business acumen has already given us great results – data scientists are able to identify profitable business opportunities and proactively go to the business with the ideas to solve it, rather than waiting for the business to come to them.
AIM: What are the most significant challenges you face being at the forefront in analytics space?
Ashish Singru: One of the most significant challenges, from my experience, is to make sure analytics is moving at the pace of business, maybe slightly ahead. If you fall behind or go too far ahead, you lose your ability to influence the business using data-driven insights. So calibrating how the analytics organisation syncs up with the business flow is critical and requires skill and effort.
The other challenge is the right balance of tech-driven and human-driven analytics work you do in the organisation – you need both and both have their benefits and limitations, but the balance depends on the data/tech maturity of the organisation and its business model. This also has implication for the kind of talent you hire for analytics – i.e. more of Business analysts versus pure data scientists.
AIM: What kind of knowledge and skill-sets do you look for, while recruiting your workforce?
Ashish Singru: The number one attribute we look for is curiosity, the hunger to solve hard problems, and enjoying the process of discovery by going deep into data. Second, we look for good collaboration and interpersonal skills – unlike what some may think, analytics careers require strong skills in these areas and working in isolation rarely results in desired impact. Third, strong database, Excel and data visualisation skills are important, though we try not to look only for a specific application or tool as knowledge can transfer from one application to another. As we go ahead, we are seeding our team with more data science talent, so that they are the early adopters of these techniques and teach others in the team as well!
AIM: How do you think ‘Analytics’ as an industry is evolving today? Could you tell us the most important contemporary trends that you see emerging in the present analytics space across the globe?
Ashish Singru: Analytics industry is currently going through some growing pains – similar to how teenagers are trying to figure out their identity! Is it a business function or technology function, which department does it sit, does centralised model work better than decentralised model. For third-party analytics firms, they are trying to figure out staffing models (project-based on resource based) and pricing models so lot of growing pains there as well. Finally, when it comes to the analytics professional, they are trying to figure out if this is an area they want to go deep into, or use as stepping stone to other functions in the organisation. They are also trying to figure out what should be their primary skill as a professional – is it more on Business analytics side, or AI/ML side or even Data Management side. It will be interesting to see how things evolve – most likely, multiple models will exist in harmony, I think.
Try deep learning using MATLAB