How technology enables operations at an online retailer

Anubhav Sudha, VP- Engineering, Hoopos

Operations for an online retailer are all about personalization, quickness and responsiveness. Any successful system for online retail would have these traits ingrained into it. In other words, any system should be evaluated for these design qualities before being considered for an online retail operation. Below I elaborate these aspects in a little more detail:

Personalized service

The shelf-space for an online store is small as compared to an offline store. The number of products that can be showcased on an online portal is much less than ideal. So it becomes important that the online portal shows clearly what is relevant to each individual customer.

The relevance applies not only to the products but also to prices, product details, recommendations and reviews. If potential customers have a specific product in mind, they would like to directly find the product using the search option on the portal. This requires an excellent search to be part of the platform.

However more often than not customers have little or no idea about what they want to buy and would want to browse and figure out what they like. It becomes the system’s responsibility therefore to assign the user to the right cluster and quickly adapt itself as if it was designed and implemented for each individual customer.

An algorithm that can quickly and most accurately identify the needs of the potential customer browsing the online store & display products accordingly would have a much better conversion rate than otherwise. In addition, it should be smart to remember the preferences of an existing customer, such as address information, payment or shipping preferences and payment mode among various other things.

Speed of service

Once an order is placed on the online store; it’s all about getting the right product/service to the customer in the promised condition and within the promised time.

A good decision system needs to be in place that can automatically decide the priority of an order, know exactly where a particular product is stored in the warehouse, accommodate special customer requests [such as gift-wrap, single/multiple shipments, courier preference] and even allow alterations to the order up to the point it is shipped. Though these examples are for an online store selling physical products, other problems of similar complexity would exist for a digital products or services online store as well.

The underlying principle is that technology should enable running the “service line” by itself with gates at the right places for intervention or manual decision-making when absolutely required. This is especially true given the huge volumes an online store has to deal with. In an offline store, lot of the decision making can be done at the POS counter and instant gratification occurs if solved.

An online store on the other hand needs supporting infrastructure to handle a large volume of orders with little or no human interaction yet allowing allowance for variability when required. To describe it in a single word, a good online retail system should be “nimble” enough to allow handling of exceptions but within the constraints of the underlying data/entity model with the ultimate goal to get the product/service in the hands of the customer as fast as possible.


The responsiveness measure of the system shows how well it behaves in times of sudden changes to the environment. Inventory management, product range, warehouse space are some of the key challenges for an online retailer. The input variables to some of these decisions change frequently at a weekly or sometimes even daily basis for online stores. A quick feedback loop is needed that can adjust the input parameters and improve the quality of decision taken.

The proposition an online store offers to its customer are convenience, better perceived value, access to better product information and a wider product range not available otherwise. On the negative side, customers need to wait for the products to be delivered and have to compromise on the desire to own the product instantly. So it becomes important that customers are informed about the status of their order at all times.

A good system keeps the customer informed about every step in the order servicing workflow and sometime even when there is no change to comfort the customer that the work is in progress. In scenarios where customer needs more information, the customer service representative (CSR) should be able to respond as if the customer was physically there five minutes ago and just came back to get something resolved.

This means that the CSR should know all the details of the customer’s transaction, any potential problem that she could be facing and be ready with the solution with a single goal of first-time successful resolution in mind. Any system which can help the CSR achieve that is a much more responsive system.

I have experienced systems getting designed either around the flow of data, products or cash or sometimes around the checks and balances or individual roles and responsibilities we want in the system. My personal view, however, is to design them around these principles and everything else should automatically become clear and fall in place.

What I mean by this is that we should not answer the question “What should the system look like for a CSR, a packer or a quality controller?” Instead we should be answering the question “Am I satisfying the need of my customer?” or “How will this help me reduce my service time?” Of course, a system cannot achieve all this in one go because no one can envision every scenario beforehand and design the system accordingly. It is an iterative and learning process which helps the system mature and finally if it is successful it would have these qualities.

Leave a Reply