#In50Hrs - How We Did It - Team Source It

This post is of how went about designing "Source It" - a local product search engine in 50 hours.

The Idea

The event started on Friday evening with presenting a rough idea to participants, in not more than 3 minutes. The idea was to provide a platform to immediate shoppers with capability to search the inventory of local shops and provide the user with best options.

The first impression I had of the idea - it was overwhelming. Specially in India, there are hardly any shops in non-branded segment with computerized inventory system. In such a scenario sourcing the inventory from this large segment is not possible. This is where we adapted our idea and narrowed it to shops which maintained a computerized inventory system.

Retail Side Validation

The first thing we did on Saturday morning was to validate the assumption from the retailers side. All of the team members individually brainstormed on the potential questions we could ask the retailers and which were relevant to our product. The list of questions we came up with were -

1. Where do your customers come from ? Do they live near the stores or come from far places ?
2. How do you reach your local customers ? Ads, banners etc. ?
3. What kind of research do customers often do before they come to stores ?
4. Is there any difference between online and in-stores product catalog ? Is there price difference b/w online and in-stores ?
5. Is online sales affecting in-store sales ? What are the strategies to counter it ?
6. What kind of location based marketing strategies used ?

Field Survey

Next, onto the field we interviewed 5-6 retailers in Koramangala area with average time spent with each retailer about 15 mins. We got very good insights of how they worked, and even got a clearer view of what our product was. Some of the insights generated were -

1. For an international sports brand company, the way they operate is through a franchise model of stores. Most of advertising is handled by the company rather than franchise itself. They are obligated to sell at the printed MRP, and does not have any leeway in giving discount.
2. The inventory data is only provided through partnership at company level and there are very few instance of that happening.
3. The mode of local advertising for most companies is through window advertisement, banners and yellow page service like Just Dial.
4. A yellow page service sends the leads of customer along with very generic information. For e.g. they would message the sport stores in a neighborhood about a customer looking for shoes, but will not be very specific about what type, brand, price range etc.
5. If a customer walks in with a particular product in mind, 80% of the sales is deemed done by the staff. They never let go of such customers and procure the product from other stores if they don't have item in their inventory.

Through the above data, we realized a user who has very specific needs is of very high value to retailers, as well as our product differs from any yellow page service by serving very specific needs of our users.

The inventory is still the toughest nut to track. At present the approach is to tackle it with appropriate organic growth strategy where stores in a specific neighborhood realizes the benefits of such a service and pro-actively integrates with it rather than separate negotiation with each of them.

User Side Validation

Problem of a multi-sided market and playing a broker between parties is you have to juggle two balls, on one hand you need to have users to convince the retailers of value they are going to get by adopting the platform, on other hand users are not going to use the platform unless there are host of options available.

Next we interacted with our users who are going to use this application. The list of questions we came up for our users were -

1. What would you like to buy locally ?
2. Do you own a smart phone ? How do you use your smart phone for shopping ?
3. How do you find local shops ? How do you find shop when in urgency ?
4. How do you find what you are looking for ? How many shops you hop to get a particular thing ? 
5. Do you check prices online before checking out local markets ?

The response was overwhelming positive and gave us confidence that we were solving a real customer problem and not chasing a fancy day-dream. Based on the response, and to communicate this user problem for the demo, we chose a novel way by creating a video.

Thanks to Shayon for volunteering as our customer.

Prototype

Since the application was supposed to be used by the user on the move, finding products in their neighborhood, it made sense to create the prototype mobile application. We chose Android because of our familiarity with the platform.

There were many learnings creating a prototype. To start with, we overcomplicated what the prototype was going to do and were soon bogged down by solving what could have been easily avoided. For e.g. we tried to have a database of products and create multiple tables to support the search function. After Yak Shaving for a good time on Saturday, on fresh Sunday morning we realized we could maintain the product data in code itself rather than database. As well as storing all the specific data, we could rather generate random data on the fly like distance of store from the current location.

Source It - Search Screen

Source It - Search Results
Displaying Stores with Inventory

Thanks to the designer in our team, the product looks very slick and professional.

Demo

Playing the ThoughtWorks University trainer role for good part of 2012, I already knew the importance of doing dry runs before the demo. As a team we decided on what roles each of us are going to play, practice our lines as well as made sure each of us were on same page regarding the product capabilities and vision.
The best part of the whole 2 days of effort was getting precise feedback from jury. The feedback we got was -

1. There is no doubt this is a real problem, and if we are able to crack it somehow, it is going to be big.
2. Cracking the inventory puzzle is going to be the crux of this problem.
3. Shiv Kumar from Exotel shared with us milo.com, a similar product in US is doing something very similar. This was an eye-opener since we were not aware of it. Checking out milo.com we realized how improvement can be made to our solution.

The code for prototype can be found on Github.

Source It Team

Overall it was an amazing experience and I hope this experience report helps others and encourages more participation in such events.