The project goal was to provide improvements for Home Depot's product search.
I needed to come up
with a high-level user flow and supporting wireframes.
Wireframing is a way to design a website service at the structural level. A wireframe is commonly used to lay out content and functionality on a page which takes into account user needs and user journeys. Wireframes are used early in the development process to establish the basic structure of a page before visual design and content is added.
See this project in FigmaThis is a demo project aimed to demonstrate my design approach to user-centered product optimizations. It took me 4 days.
Conduct user experience evaluation of Home Depot's product search subsystem—and provide solutions on how to make it better.
Stakes of e-commerce search are exceptionally high: if users can’t find the product they’re searching for, they can’t buy it. This not only leads to lost sales here and now, but it also makes users less likely to return in the future as they obviously won’t go to a store they don’t expect to carry the product they are looking for.
Search interface and behind logic need to ensure users' ability to find relevant products. But in accordance with the Baymard Institute usability study, even multi-million dollar sites had so poor support that 31% of all product finding tasks ended in vain when the test subjects tried using search.
I started my work from the investigation of public information: the brand history, product categories, target markets, audience segmentation and interaction channels.
One of the key things a solution designer need to understand when on a project is what problem(s) need to get solved.
That means you should know your users, and understand their specifics and needs. Therefore I started problem research from the definition of high-level needs for the defined key customer segments.
Guerrilla testing is one of the simplest and cheapest form of user testing. Using guerrilla testing usually means going outside to ask people there about your product or prototype. It can be conducted anywhere ex- cafe, library, train station etc, essentially anywhere where you can find a relevant audience.
I conducted a few moderated think-aloud usability studies with my friends. To finish this project in time, I focused my case study on desktop visits, most of the findings are applicable to the mobile user experience as well.
I investigated what users expect as they perform product searches on Home Depot's web site, their search strategies, what typically goes wrong in the process, why it goes wrong, and exactly what changes it will take to avoid these issues.
I created a Customer Journey Map for the selected user segment: Do-It-Yourself buyers. It helped me to summarise the user study insights, and to figure out optimizations are necessary to ensure users' ability to find relevant products.
Then I needed to define key search strategies, search subsystem elements and search logic necessary for this particular e-store.
I didn't ask people to search using the global header search form. Thus, they were completely free to choose whatever search strategy they wanted (including category navigation, search engines and venturing off-site). A large portion of them consistently used site search subsystem as a primary product finding strategy.
Heuristic Evaluation is a detailed analysis of a product or its components that highlights good and bad design practices in existing implementation. It helps UX designers learn the current state of the subject in terms of usability, accessibility, and effectiveness of the experience.
I analyzed the Home Depot's Search Subsystem using best practices from Baymard Institute research, good solutions that I noticed during the competitors review, guerrilla testing, own experience and brand understanding.
1) Some users expect search to include all content on the site – not just products but also auxiliary content such as help pages and store information.
2) Users often inject subjective components (quality, beauty, value, etc.) into their search queries that require the search engine to go beyond accuracy and venture into interpretation and opinion.
3) Some users write search queries the way they would explain them to someone else (i.e., as a full spoken sentence), yet many search engines have difficulties interpreting these natural language queries.
4) Sorting site-wide search results often leads to highly irrelevant results ranking first.
5) Accessory products often pollute search results and render price sorting useless.
6) The layout of the search results can severely limit the user if not properly matched with the product type (list-view vs. grid-view).
7) Existed “no results” search page is unhelpful. It does not guide the user back on track as it contains an only category navigation menu.
To communicate suggested UX optimizations I created medium-fidelity prototype for a search results page layout and related interactions.
When e-commerce search works, it can feel almost magical: you simply type in what you’re looking for and it is served up in mere milliseconds. It’s fast, convenient, and super-efficient. But unfortunately, e-commerce search all too often doesn’t work that well.