HUMAN-CENTERED RESEARCH INTO THE WORLD OF RETAIL PURCHASING

As part of my Master's in Human-Computer Interaction program at Carnegie Mellon University, the client for my capstone project approached my team of 5 with the goal of designing a new collaboration platform for retail buyers and vendors. To begin the design process, we conducted 4 months of human-centered research in the retail buying field, using the methods summarized here. We're now iterating upon the design of a web-based tool (so check back again soon!).

Client:    LH Ventures, a start-up/venture capital firm

Audience:    Retail buyers and vendors, especially at off-price chains

My Role:    Documentation Lead, UX research

Duration:    Jan 2016 - present


User Groups

 
 

Challenge

How did that new shirt you just bought make its way to the store? Every day, retail buyers make decisions that determine what products are available to you, the consumer. The cumulative decisions of a retail buyer affects the success of their company. However, their decision-making process is not wholly data-driven, and their communication with vendors is inefficient. LH Ventures asked our team to design a new SAS software that would aid both retail buyers and vendors. 

 

Solution

Over a period of 4 months, our team conducted over a hundred hours of user-centered research to understand the motivations and processes of buyers and the stakeholders influencing them.  We identified opportunity spaces for a new tool to help both buyers and vendors get their job done more efficiently, without sacrificing the thrill of a good deal. We're using these findings to guide us as we iteratively build prototypes in the summer semester. 



Research Methods

 

In order to begin designing a tool for retail buyers, we asked ourselves a lot of questions, including...

Question #1: What is retail buying???

 

We started to learn the lingo through literature reviews.

POS, FOB, POE, keystoning...retail buying is a jargon-heavy field. We began our exploration by learning the key terms, even creating a matching game to test ourselves. We then looked at over 26 papers and articles on the field of retail buying. We learned how buyers make their purchasing decisions, creating this diagram to summarize our insights.

 

We used stakeholder mapping to learn who was involved and their roles.

Now on stable ground, we worked with a Tepper Business School subject matter expert to outline all the stakeholders directly or indirectly involved in the retail buying process. We detailed their tasks and relationships, and began feeling comfortable in this new field.

 

We worked our butts off recruiting...and were able to interview over 20 buyers and vendors. 

We conducted 22 interviews with individuals from all over the retail buying map, both remotely and in-person. These interviews were primarily semi-structured, with a common core of questions. However, because interviewees came from many different roles, the interviews had to be tailored to the individual. 

 

 

 

We observed 6 buyer-vendor meetings using contextual inquiry techniques.

One of the most exciting and informative parts of our research was sitting in on sales meetings in the offices of several major retail chains. It was eye-opening to get behind-the-scenes of how products get chosen. We used affinity diagramming to record the rich information we gathered about work practices; the social, technical, and physical environments; and user tools. 

 

...and then we synthesized our learnings into a flow model and several personas. 


Question #2: What is a retail buyer's typical workflow?

 

We asked buyers to detail and rank their daily tasks using a card sort. 

We asked our card sort participants to first brainstorm a list of their daily work tasks. From this list, participants were asked to arrange these tasks along a number of axes: unenjoyable to enjoyable, no data required to a lot of detail required, individual to in-group, inefficient to efficient and more. We saw opportunities in tasks they rated as inefficient, and made sure to note to keep tasks they found enjoyable (interestingly, sometimes one and the same).

 

We gathered information about 20 current tools through a competitive analysis. 

Since we were designing a new tool, we also knew we needed to give our clients a competitive edge -- understanding the current retail software landscape was vital to helping our client develop and carry out their business model. We attended demos of tools and scoured online reviews and videos.

 

We analyzed over 50 documents and emails generated from the buying process.

We maintained a thick binder of representative documents. By analyzing them, our team was able to identify and categorize additional, and more specific, breakdowns in the flow of information between buyers and vendors. 

 

...and we synthesized our new understanding into process and customer journey maps.


Question #3: What makes a retail buyer tick?

 

We asked buyers' opinions through an online survey.

In order to validate our understanding of buyer preferences and experiences, we created and distributed a survey. This survey provides more quantitative support for our insights and models. The survey was distributed through a number of retail buyer LinkedIn groups, and directly to some of our contacts.

 

We put ourselves in their shoes through an empathy-building activity.

We wanted to gain firsthand experience with the difficulties that retail buyers face everyday—trying to understand their customer and making  good purchasing decisions based on relatively little data. So we created Duè Shop: a shop where we sold baked goods based on forecasted consumer data. We experienced just how difficult making decisions as a buyer can be, but also how rewarding it can be when you make a profit!

 

...and we grouped our observations into a cultural model and affinity diagram.

 

Question #4: How can we satisfy retail vendors' needs too?

 

It takes two to tango for any collaborative platform. Throughout our research process, we sought to understand the vendors' perspective towards and needs within a sales platform. We used the following Venn diagram to summarize the overlap between the two parties' values, documents, and tasks -- the sweet spot within which to develop an effective tool. 


We began to envision solutions.

 

We capped off our Spring semester of research by creating 3 assumption prototypes, quick-and-dirty physical manifestations of a few our solution ideas.


And we used the next 3 months to iterate on the design.

 

Our Capstone project continued with the Summer semester, devoted to iterative design, with low- to high-fidelity prototyping and user testing.

 

LESSONS LEARNED

 
 

Lesson #1: Recruit early and often

We were mostly on our own for getting participants for a very small niche of user types. After a slow start recruiting, we learned our lesson and recruited as often as we could. And we got pretty creative doing it: contacting buyers and vendors through LinkedIn, Facebook, Twitter, and even--I kid you not--through YouTube. 

Lesson #2: Make the textbook user research methods your own.

One of the main purposes of the Capstone project is to apply recently-learned user research methods in a real-world setting. My team found that as we became more comfortable with the research methods, we found we could riff on them to create our own method variations and hybrids. For example: we used our own color-coding system within our affinity diagram to combine participant quotes from multiple roles. 

Lesson #3: Group synergy is elusive and hard work, but oh so sweet.

Rallying together five almost-strangers to work together for hours on end, accomplish a goal, and still be friends at the end is tough, no doubt about it, especially when there were no delineated roles or responsibilities to begin with. But I found that when we got into a group groove there was nothing more productive or exciting.

 

Note: This content is Proprietary and Confidential Information of LH Ventures, published with their permission.