Airtasker Case Study
PERSONAL PROJECT
The challenge
Before joining Airtasker as a Product Designer in June 2018, I undertook a design challenge. This challenge is no longer run internally so I would like to showcase the submission. Here was the brief:
Redesign the My Tasks section on the iOS or Android app
In my process, it is always important to understand the ‘why’ behind any task. By defining that answer, the project will have a north star to guide any decisions.
Why did the ‘My Tasks’ tab need redesigning?
What are the strengths and weaknesses of the existing design?
What areas were users struggling with when using this part of the app?
Were they even using it at all?
Below is an outline of how I approached this challenge, and what I believe is a solution that would increase engagement within this section of the iOS app.
Discovery
OBSERVATION
While the brief specifically mentioned the My Task tab, it was important for my research that I kept a broad focus in the initial phase.
I wanted to understand how existing users use the Airtasker app on a daily basis. To do this I performed a mix of face to face and remote interviews with both Taskers and Posters. The goals of the interviews was to understand the perceived usefulness of the My Task tab, user’s understanding of the taxonomy used as well as the general sentiment towards Airtasker as a product.
BUILDING EMPATHY
What better way to walk a mile in the customer’s shoes, than to actually become an Airtasker member?
My front yard was in much need of some love and attention, so up went my first task. I also dabbled with making offers on tasks and completing them so I could gain an insider view on the end to end process as both a Tasker and Poster.
TERMINOLOGY
Whilst using the app myself, I felt a lot of confusion around the terminology used in the filtering on the My Task page. I wanted to discover if I was the only one, so asked users were to describe what they would expect to see for each filter type.
Generally users understood what each meant, with the exception of the Projects tab. Click on the button for a breakdown of the test results.
“It’s all a bit unclear. You don’t know if the offer pending is talking about you, or them posting it. That’s why I like when it says all tasks, cause you know what’s going on. If it just said ‘posted tasks’ and ‘accepted tasks’ it would make more sense. I wouldn’t have five options, i would keep it to three.”
Customer feedback from the Apple App store.
User Interviews
Interviews were conducted with:
Jessica Cuda, 27yo female, Mascot, Photographer, daily tasker.
Sarah Hewington, 19yo female, Strathfield, Student, daily tasker.
Deborah Dean, 28 yo female, Albion Park, Casual employee, daily tasker.
Justin Murray, 29yo male, Sydney, Systems Admin, first time user.
Ela Highland, 33yo female, Paramatta, Doctor, daily poster.
Heinz Neubauer, 36yo male, North Shore, Project Manager, casual tasker.
Vedha Nayak, 31yo female, Mascot, Web Designer, casual tasker.
Kenny P, 23yo male, Yagoona, Linguist, first time user.
Trac C, male, Strathfield, Landscaper, daily tasker.
Four of the interviews with Taskers were one hour each, allowing me to deep dive into the following user goals:
Finding suitable tasks
Make a winning offer on a potential task
Scheduling work and ongoing communication with Posters
Monitoring the progression of existing tasks
Monitor the status of tasks that offers have been made on
Get paid for the work completed
From here I grouped similar pain points onto a high level journey map, as well as a separate map of identified opportunities. At first it looked like the biggest frustrations hung around the quality of the tasks being posted.
“[The My Task tab is] not something I use. It doesn’t make a heap of sense. I don’t understand why it’s there. It seems useful, but convoluted.”
Frustrations
Users are not aware that the filter tab exists. Convoluted workarounds exist because users haven’t discovered how to use the My Task filters to their advantage, specifically the ‘Offers Pending’ filter.
The majority of interviewees prefer to take communication outside of the Airtasker app. This is due to speed and it removes restrictions experienced on the app such as attaching large files.
The purpose of the page is unclear to first time users. The ‘empty state’ is not being utilised to its full advantage, there is potential to educate new users on the terminology used within the app, and push users down new paths instead of currently ending their journey.
The taxonomy ‘Projects’ confused all interviewees. They had no idea what to expect on this filter, or what the term was referring too. After applying the filter, their confusion was not relieved. Some stated that it’s purpose was a duplication of other filters.
Users are unclear on whether the tasks shown are from a Tasker or Poster point of view.
Message signals are unreliable, and often missed. Has affected jobs before when the Poster has missed message alerts.
Users generally aren’t aware they should be declaring their income, and use the My Task tab to keep track of their income. This is problematic because they do not move on to discover the Payment History feature.
Posters who stop responding after two days are disheartening and usually a red flag for Taskers to abandon interest in the Task. “If the person doesn’t respond to questions, do they really want it done? Like why bother? 4 to 5 days is too long, if they’re not answering questions, they are going to be hard to work with later.”
The number of filter options is too high. Offers accepted and offers pending could be combined to be simplified.
OPPORTUNITIES
When asking the Taskers what features they would like to see on the app to make their lives easier, most items were related to the pricing of tasks. Items ranged from the 15% Airtasker fee, more realistic pricing from Posters through to improved on boarding.
The other filters, while difficult to locate, made sense and were interpreted correctly.
The mobile notifications drive most of the interactions within the app.
The My Tasks tab is useful for scheduling. Would be more useful if it could show deadline, potentially explore a GANTT chart style system. “If i see the date and it’s due on assigned tasks, helps prompt to get ahead.” - Sarah.
Users move accepted tasks into their Google Calendar when details have been confirmed. Would be useful if it synced.
Overdue is really good to keep track of your jobs and the amount you got paid. Helped to follow up on unpaid jobs. Uses it to track ‘where you’re at’ and to see who has paid you and hasn’t.
Define
From the research, the next step was to define the problem statement for the project. I focused my efforts on synthesising the information that was directly related to the My Task tab. From there, I refined the multiple pain points users were experiencing, which lead me to the following statement:
Users are failing to understand and engage with the benefits of the My Task tab due to the ambiguous interface.
Assumptions:
Users don’t use the filtering on the My Task page
Users aren’t reaping the benefits of the My Task page
Users don’t understand the purpose of the My Task page
The My Task page’s interface is ambiguous
Identified pain points grouped by theme.
While the above pain points are outside of the scope of the 'My Task' tab it is important to acknowledge other areas where the user experience can be improved. Below is a list of key themes from my research, note these are from a Taskers' point of view.
Poor quality posts
Large volumes of irrelevant tasks
Lots of ignorance and naiverty when it comes to pricing
Abuse of the market place with degrading pricing
Low profits to be made
Communication
Competition for the quality posted tasks
Poor initial onboarding
Ideate
Once the problem statement was defined and validated through interviews and feedback, it was time to theorise some possible solutions that would help acheive the projects goal.
How might we encourage users to discover and utilise the benefits of the My Task tab?
I believe if we introduce a Poster/Tasker toggle, users will no longer be confused by which task belongs to who, because a distinction is clearly defined.
I believe if we change the styling of the filter icon to be consistent with other pages, engagement will increase, because users will be more likely to discover it exists.
I believe if we let users filter their offers by the passed response time from users, they will stop performing cumbersome workarounds because they will use the My Tasks tab instead.
I believe if we introduce sorting to the My Task tab, more users will interact with the tab because they have more control and flexibility on how tasks are displayed to them.
I believe if we add a CTA to the empty state, along with custom text, page depth will increase because first time users will be more educated and will follow the CTA.
I believe if we change the terminology used for a task’s status, Posters will pay their Taskers faster, because they are more aware the task is unpaid within the app.
I believe if we introduce message notifications within the My Task tab, users will engage with the page more because the information presented is more useful.
Wireframing
There were a few different ways I could execute the changes to the filtering. I also wanted to introduce sorting. Different UI patterns existed for both sorting and filtering within the iOS app. For consistency I felt that it was important to repeat the existing patterns and leverage the learned behavior so that the new design changes would cause minimal friction.
I also made a call on which display methods would offer the most flexibility. For example when selecting which status types to appear, it would be straightforward to mimic the existing sort dropdown. However users had stated a preference to view multiple status types, so a checkbox list would give them that versatility.
Rough plan for new task flow and terminology.
Updated task status
From the research, it was clear that the meaning of some of the terms used for task status was not clear. Jessica has told us her struggles with having to teach her clients (Posters) how to use the app, as first-time users had a pattern of misunderstanding that they needed to release the funds in order for Jessica to actually be paid.
I also had some assumptions that the phases of the task (open > assigned > completed > reviewed) lent to the confusion. I wanted to see if introducing more status, as well as rewording them to be focused on the outcome required from the user, would help alleviate some of the frustrations both user types were facing.
Iterative Design and Testing
I create a full visual mockup of the Tasker page in it’s empty state as well as populated with a few jobs. I made sure to include the new status changes as well as my initial preference for menu behaviour.
Before: Original load state
After: Design round one, load state
User Testing Round One
These images were tested on Veda and Heinz, a Tasker and Poster. Initially I wanted to run the static images by them, so restricted this testing to first impression tests to gauge if I was on the right path. I also asked some more questions focusing on the expectations around filtering and sorting, and how they would envision this working.
Heinz
Noticed could toggle between tasks. “I like the toggle because before all of the columns were mushed together. I have no interest in posting tasks, so this is super convenient.”
Assumed if clicking the filter icon, he would be able to filter as per previous function.
Saw a clear difference between offered and assigned tasks. Liked that they are prominent.
Quickly noticed the ‘new’ for messages. Didn’t use the filters much on My Task tab.
For tasks that offers have been made on, but didn’t win, Heinz doesn’t expect them to stay on the task tab. There is no need to know that it has gone to someone else, otherwise it could grow to a massive list. It would make sense for it to disappear, but to also have some kind of indication that he didn’t get the opportunity. But shouldn’t be in the main section.
Would expect the filter to show check boxes so he could select which status to view. After being showed the existing filtering “I don’t like that. Naturally I wanna see more. It would really annoy me [if i could only select one status at a time].”
Vedha
She would like to see the completed tasks separately.
Would still want to see the tasks that have been declined.
Stated importance of being able to filter, as found the mixed status types misleading.
Had not noticed the filter button in the interface.
She stated that it would need to be explicitly the difference between the two, and went on to suggest tabs. This suggested that she initially failed to notice the toggle at the top of the page.
Co-design sketch-up
After initial questions, we then did some co-design sketch ups to envision how they would want the filtering to work. Both users had very different opinions, so further testing is required.
“In the beginning it is useless. I’d just have it set to all. The number of my tasks is so small, so no need to filter. I probably want to see Assigned and Offered, but I wouldn’t want to see completed because it has little relevance to me. I’d like to check it so it’s filtered out from view.”
Learnings
Add ‘Declined’ status to task workflow
Existing users may struggle with ‘finished’ due to it’s similarity to ‘completed’
‘Unreviewed’ can be interpreted as the Poster needing to review the quality of work completed before releasing payment. This is the opposite intent.
Still need to validate the desirability of the date range filter.
Still need to validate that the majority will find it useful to segregate pending offers that have low response rate from a Poster.
There are mixed opinions on what rules to use for sorting tasks and which default method is best.
Clickable prototype version one
User testing round two
Any actionable feedback was implemented and more screens mocked up to reproduce the flow of using both sorting and filtering within the app. I used Invision to make the designs into a clickable prototype and then ran some final tests to validate the changes.
My final user for testing was Kenny, who had not heard of Airtasker prior to this test.
Findings
Quick to interact with the sorting feature, however the expectations were not a match. “I want to immediately filter, i’m not certain that the sort would do this”.
Unfortunately Kenny failed to identify the filter icon. At the end of the test when I informed him of it’s existence, he confessed he thought it was the logo for the app. I would suggest researching the recognition of the filter icon for new and existing users, in both the context of the My Task and Browse page.
When shown the filtering he had this to say: “It i only wanted to see one thing at a time, it’s quite cumbersome. I do like the functionality, however it would take longer to do what I want it to do.” When shown a design for the alternate option of having a drop down style selector he said he preferred the toggles, as he had the choice to view multiple at the same time. He would just like to be able to additionally select ‘this only’ so that one status could be viewed at a time without having to deselect the other options one by one.
I had also introduced additional filter options which Kenny was able to correctly identify the purpose of.
The ‘overdue’ status still caused an issue, so for the next iteration I will replace with ‘missed’.
As there is still division against sorting and filtering, I would recommend performing tests with a larger pool size, focusing on understanding what information users value most. And from there determine the preferred method of displaying and filtering.
Measure
As there are quite a number of changes proposed, I would work on testing each hypothesis individually, by releasing multiple A/B tests to a small number of users. I believe that major design decisions should be driven by data. While qualitative research helped identify the original pain points, quantitative data from analytics would help validate some of the assumptions.
I would also ensure that data is being captured in relation to our goals, allowing us to measure the overall success of the project.
Feedback
“I like how you have separated as a tasker and as a poster. Looks good. And the new in red is good too.”
“I wish that was real airtasker! I like the different colours for the unpaid and overdue signs. It’s very clear as a poster and a tasker.”