How to pull together quantitative data on user behaviors and mental models
Later than getting really beneficial insights from our before yap usability test, we wanted to explore the in sequence architecture featuring in greater facet.
So we ran a first-click test on Yelp’s homepage using Chalkmark, which gave us insights into the aspects of the in sequence architecture with the aim of worked, and which aspects befuddled users. And we ran a congested tag sort on Yelp’s be included search filters using OptimalSort to obtain old hat which search filters were the the largest part well-liked, and if near were slightly filters with the aim of may well ensue disinterested to reduce clutter.
The quantitative methods we used were all count and cost-efficient, demonstrating with the aim of user make inquiries doesn’t require thousands of dollars, a team of researchers, and endless count.
Such as discussed featuring in The escort to Usability taxing, apart from of the method you pick out, recently remember with the aim of user make inquiries is not going on for characters reports — it’s going on for asking and answering the precisely questions and gathering data so with the aim of you can cause evidence-based decisions featuring in your designs.
Choosing our users
We recruited participants through Optimal Workshop’s recruitment panel, and reserved our demographic to citizens based featuring in the US (as with the aim of is someplace yap is widely used).
As a replacement for of filtering by age, gender, earnings, or else central processing unit experience featuring in the recruitment stage, we asked pre-activity questions designed to expand our understanding of participant responses. This is for the reason that while demographics are of great magnitude, come again? Users know and how they avail yourself of analogous products is likely new of great magnitude.
Meant for the congested tag sort on Yelp’s homepage, we asked participants how often they used yap, and how often they used the search filters, so with the aim of we may well filter our data based on users with new or else a smaller amount experience. Knowing how many citizens used search filters would besides present us an insight into how of great magnitude search filters in point of fact are meant for Yelp’s users.
Meant for the first-click test, we asked participants how often they used yap, and their likelihood and frequency of characters yap reviews. We wanted to know how often they wrote reviews so with the aim of we may well verify their level of comfort with the situate.
Since we sought quantitative data, we wanted to recruit a lowest amount of 30 participants meant for both study (NNGroup Principal Jakob Nielsen recommends a lowest amount of 20). We ended up taxing 40 citizens meant for the congested tag sort, and 38 meant for the first-click test.
If you’d like to realize new going on for screening and recruiting users, check old hat the NNGroup’s of use and on the house e-book with 234 tips.
We ran a congested tag sort using OptimalSort
Congested tag categorization involves presenting participants with labelled cards, and asking them to situate them into pre-defined categories. Such as Donna Spencer says, congested tag categorization is a ‘content-centric technique’ and can ensue beneficial ‘when toting up in mint condition content to an existing constitution.’
Once it comes to naming the cards, simpler is better. Elude substantial lexis (many syllables) and jargon. This advice is essential meant for tag categorization since unnecessarily hang-up labeling preference disrupt natural notions processes.
Pierre Croft, IA and UX expert meant for Decibel Digital believes with the aim of tag categorization can even help defend next to the bad ideas of HIPPOS (highest paid citizens featuring in the room) who generally aren’t the muddle usability experts. Tag categorization is on sale, beneficial, and quick, so we’ve integrated a little pointers which apply to congested and unlocked tag categorization:
Don’t mix mother and toddler categories — featuring in other lexis, avail yourself of categories from the same level, or else to boot you preference confuse your participants.
Provide unlocked forms meant for other advice later than the test— While this is standard procedure meant for unlocked tag categorization, it’s besides quite beneficial meant for congested tag categorization. Provide a combine gap forms (or gap cards) meant for participants to mark down other categories. While the in sequence might ensue “off-the-record,” it may well bring beneficial insights.
Don’t intervene too much— later than giving the commands, try your greatest to recently sit back. Intervention can mask the data. Luckily, this is not an installment featuring in remote tag categorization.
Recognize with the aim of every now and then users don’t convene everything — A lack of grouping can ensue recently such as illuminating such as a structured categorization. If this happens, cause certainly you ask the user why. If you’re running a congested sort and not everything is sorted, you can besides provide gap forms (or gap cards) to go with why the existing categories weren’t chosen.
As a replacement for of taxing the top level navigation labels of Yelp’s website, we determined to avail yourself of congested tag categorization to obtain old hat which facial appearance search filters were the largest part of great magnitude to users, and which were disregarded. This analysis might help simplify the search filter options, such as ‘visual clutter’ was mentioned by participants such as an installment.
Our congested tag sort had three clear-cut objectives:
Determine how often citizens avail yourself of search filters on yap (or a analogous site)
Determine which filters are the largest part of great magnitude to users
Determine which filters are smallest amount of great magnitude to users
Featuring in calculate, we had 47 cards representing all of Yelp’s 47 search filters (price, distance, etc). We therefore asked participants to sort them into categories of significance: Very of great magnitude, somewhat of great magnitude, not of great magnitude, and unsure.
A first-click test using Chalkmark
First-click taxing records the users’ pioneer click featuring in response to a task. Participants solitary click formerly, and therefore move on to the subsequently task. First-click taxing is seemly increasingly of great magnitude: Studies hold revealed with the aim of if a user gets their pioneer click precisely, they’re 87 percent likely to complete the task they came to the website to complete.
First-click taxing can ensue ended on a live website, initial prototype, or else even recently a wireframe. Jeff Sauro, Founding Principal of MeasuringU, recommends conducting first-click taxing later than both key iteration. At this time are a few guidelines to get the gist:
Mark lucid tasks — recently like you would meant for a scripted usability test, cause certainly the participant is thinking going on for how to solve a setback as a replacement for of recently someplace to click. Facet isn’t essential, but clarity is.
Label the greatest paths to triumph — Start from the homepage and plot all likely paths with the aim of preference appropriately accomplish both task. First-click taxing is even new applicable if your situate gets a great volume of search traffic (like Yelp). For the reason that your homepage probably won’t ensue the pioneer send a message users obtain, first-click taxing ought to ideally ensue ended across your whole situate.
Count both task — A 90 percent first-click rate on the correct label might deceptively indicate with the aim of your navigation is of use, if you timed the test and adage it took an usual of three minutes to cause with the aim of pioneer click.
Our first-click test had two objectives:
Determine if the in sequence architecture enabled users to complete tasks quickly
Determine if the navigation labels are lucid
We asked users to accomplish some tasks (such such as ruling a reliable nightspot later than ceremonial dinner featuring in San Francisco), provided them screenshots of yap pages, and recorded someplace they clicked. We therefore analyzed the heatmap results, and the break the speed limit with which participants complete the tasks they were presented with.
These two remote make inquiries techniques are two with many
Such as user researchers and UX designers, you hold an almost endless add up to of techniques and tools to pick out from once you embark on a design or else revamp project. Meant for us, congested tag categorization and first-click taxing provided the greatest balance of data, cost, and break the speed limit. We knew with the aim of these techniques would provide us with quick data to support our qualitative make inquiries, and results with the aim of would ensue effortless to explore and crowd-puller design recommendations from.
Tags : OptimalSort
