Datafiniti crawls the web and compiles it into a massive database. The data is then sold via a subscription service to companies who use it to, among other things, learn competitive pricing information, find sales leads, and do reputation management. At the start of this project, Datafiniti only offered its customers access to its data via an API. To expand the company's customer base, the CEO decided a less technical and more user-friendly interface was needed.
This was the first project I did while working at Datafiniti. My role on this project was as in-house UX designer. I also acted as product manager and team lead for the small team that develops Datafiniti's web-based products. As part of my duties I conducted user research with both external customers and internal stakeholders, and created a set of design deliverables (shown below).
My research largely took the form of interviews but also included reviewing Datafiniti's competition. Datafiniti as a company was fairly new at the start of this project and had very few customers. The customers it did have used the company's API and were technical users; they likely had some programming ability and knew how to work with data in JSON format. These customers were not the target audience of the new user interface, but they could still provide valuable feedback. I also talked to a few potential customers, people who were in the market for the kind of data Datafiniti sold. I asked them about how they planned to use the data, what other data providers they had looked into, and what their ideal solution would be. Many people wanted a simple solution, something that would allow them to get data quickly and easily. After reviewing the product offerings of Datafiniti's competition, I understood why simplicity was important. Most of the competitors' interfaces were very long and complicated.
Each record in Datafiniti’s database has dozens of attributes that can be searched by. I decided showing all of the attributes in the interface would go against the requirement of keeping it simple. Using analytics run on API searches, I determined what the most searched-by attributes were for each data type, and used those as the interface's search criteria. I then created three interface designs based on the research I had gathered and reviewed them with the same group of people I had interviewed earlier.
The first interface design was based on a common phrase I heard during the user interviews: “Make it like Google.” However, using natural language search techniques was beyond the Datafiniti development team’s capability. Instead, I chose to start the user off with a small set of search criteria, and then let the user continue to search, sort, and filter the data on the subsequent results screen. The response I got when reviewing these wireframes was negative. If users didn’t initially see the criteria they wanted to search by, they assumed the UI wouldn’t let them search on it at all.
The second interface concept I designed was more complicated. The idea behind this concept was to allow users to create a database query. Users could choose a data type, then they could choose which criteria they wanted to search by, and finally enter values for the criteria. This interface concept also received fairly negative reviews from the group I reviewed it with. Some didn’t like that it started with a mostly blank screen, while others thought it was too complicated.
The last interface concept was actually the results screen from the first concept. This interface uses the standard convention of having search criteria on the left and results in the main content area. Many ecommerce sites use a similar layout, making this design more intuitive for users. This concept received the most positive response.
The design for the search interface has gone through several revisions since it was first launched. Shown here is the most recent design. The design is optimized for desktop / laptop use, as 98 percent of Datafiniti’s users are on those devices. Other parts of the interface, such as the sections that allow users to view and download result files, and to manage their account, are mobile friendly.
A complete prototype of this site can be made available on request. A demo of the live product can be done in person.
As an exercise I created a set of mock-ups for a mobile version of the Datafiniti search interface, using the Android Material UI style. This design will likely not be developed, though it will certainly inform the design of a responsive version of the web interface if the need arises to create one.