The retail sector
Mom-n-Pop stores comprise 94% of India’s retail sector. Their presence in every corner of the country helps them serve the daily needs of every Indian.
With this reach, however, has come inefficiency and miscommunication between the supply chain entities, resulting in numerous issues – most affected being these small businesses.
The business ask
At the same time, the FMCG companies were unable to track the last-mile-sales when it mattered the most. So they asked SAP if we can address this problem and help them gain insights into this very critical last mile retail problem.
To understand this problem better, the first step was to realize how big of a problem are we talking about? When we looked at the entire eco-system of the supply chain in India, we realized that the scale of this problem was humongous. 94% of retail business happens in these Mom-n-Pop stores, which in a country like India translates to about 33 Billion Dollars per year.
There are about 50 major CPG companies with a network of 20,000 distributors and 8.3 million retailers catering to 1.3 Billion consumers.
As a very first step to understanding the issue from our primary stakeholders (which happened to be the CPG companies that are existing SAP customers), we started talking to them to understand the problem in detail. Based on what we heard from them, we were able to visualize the problem as three pieces of a puzzle.
So we embarked on a journey to uncover this.
We talked to multiple CPG executives, Distributors, and Retailers. We observed them do their business and shadowed them for days.
Based on this fieldwork, we were able to create a map of issues faced by each of the players in this ecosystem. Each one of them had a unique set of problems. Most important were the things that CPG companies wanted and the problems faced by the retailers at the last mile retail.
Initial user research pointed out that this is a system design problem. We needed a way for CPG companies, Distributors, and Retailers to somehow
The distribution system behind this supply chain ecosystem was sub-optimal. This meant that retailers were receiving access stock of things they don’’t need, and none of what you actually need. We realized that by the time the retailer is able to communicate the demand to the distributor, and they to the manufacturer and they finally receive the goods, it’s already too late and their customers have moved on and the retailer has lost a valuable customer.
We designed a solution that maintains the flow of information between the retailers, distributors, and CPG companies delivering real time information grounds up.
The main finding was that all these years, the retailers have been doing their business on pen and paper. This was making it harder for CPG companies to obtain transactional data on what items were moving fast, what is in stock, or how their promotions are affecting sales.
This solution had multiple parts:
The entire solution was designed around the point of sale device.
With this, we were able to connect all the parts of the system and provide real time insights about the sales, inventory, orders, and invoicing.
After defining the flows, we started working on design concepts. A big part of the problem we were trying to solve involved dealing with huge amount of data and dynamic information that needed to be sliced and diced by the end users. So we spent time on crafting easy to use data visualizations.
Once we had the data started coming in, another big challenge ahead of us was to figure out how can we slice and dice the huge amount of data that our users to help them generate insight from this data.
The toughest one was for the CPG dashboard. At any given point, the executives should be able to review real-time data across locations, time frame and huge sets of product categories.
We explored multiple design solutions for the filters. Every time going back to the users and testing each design. With constant feedback and active involvement of the end users, we were able to quickly prototype and refine these early designs.
When we interviewed our stake holders, we realized that they are all looking for multiple ways to view this complex mesh of data so that they can make sense out of it. To understand this, we asked them a lot of questions and documented what they were saying.
Just to give an idea, these are just some of the many permutations and combinations they were thinking of:
As you can see, this gets more and more complicated.
So we created a structure to help analyze these permutations and combinations into a easy to use user experience model.
Putting it all together
After a lot of back and forth, and having done these individual studies on the key components of the interface - The Layout, Filters, Data-point visualizations etc. we started putting it all together to create a cohesive user experience.
We got a mixed response to our initial design proposal from the users. On one hand, the users loved the amount of data they can analyze in real-time and how they can mix and match different data-sets. But they were not able to understand the filters properly. They specifically didn't understand the drop-down filter. Every time they would need to open it, the rest of the user interface moved down. A lot of movement was also felt as a distractor.
Users complained that direct manipulation of category and location fields was not possible unless you opened the filter dropdown. This also resulted in two extra clicks to get the job done.
Lastly, we also got mixed reactions on the choice of colors. Most of the users felt distracted by the use of bright colors and a very strong accent color. The key information was losing focus and the user's eyes were directed more toward the action buttons.
It was time to iterate on the designs based on the feedback. We set out to re-designed the interface to make it less distracting. We also removed the animated drop down and incorporated a much cleaner version of the filter that would allow direct manipulation of filter parameters.
SAP Ganges was shortlisted at the IxDA 2015 UX design awards under the "Optimization" category.
We realized this a little late in the project. After our initial user research, we should have involved a users a bit more in the design process to understand their preferences. This would have saved us time we spent fixing some of the issues that could have been avoided.
When we are dealing with a lot of data, we tend to think of highly interactive widgets to help users navigate the data and make sense out of it. But sometimes we stretch these interactions too far and instead of making the data consumption easy, the widget themselves become complex creatures. Finding the right balance between the interactivity and simplicity was the most challenging part of this project.