Increasing candidate onboarding %

redesign • user research (b2b) • prototyping
INTRODUCTION
Tempo is a SaaS-based hiring tool, matching early career candidates to junior commercial and admin roles in the UK. The success of the product is partly driven by the quantity of available high-potential candidates to be discovered by hiring managers
MY ROLE
Research, analysis, product design, workshop facilitation and user testing
THE TEAM
1 Product Designer (me), 2 Developers, COO
TIMELINE
June 2022 - September 2022
OUTCOME
+50% onboarding conversion, leading to a higher availability of quality candidates
Problem context
We discovered that a significant proportion of new candidate users were dropping off between hitting ‘sign up’ and completing their profile
impact of problem
⬆️ Increasing cost per acquisition (CPA)
Tempo invests in various online strategies to raise awareness and attract candidates to the platform
⬇️ Decreases funnel of candidates
High volumes of fresh and high potential candidates help hiring managers hit their ‘Speed to hire’ KPI
high overview of the old experience
Note to self: Next time a project might be case study worthy - screen shot the old experience asap before it starts to get strange why you might have to ask your developer colleagues to give you a preview environment before the update rolled out
if you are wondering... "can we skip to the good part" 🎵
jump to solution highlights 🙃
impact of problem
Onboarding goal: capture sufficient profile information to generate ‘matches’
Until candidates provide the defined sufficient level of information, they are technically ‘invisible’ to employers on Tempo, as without a complete profile, hiring managers would have no basis to screen candidates for their positions. Likewise, candidates can only begin to see relevant jobs when their profiles are complete.
research methodology
Diagnose the problem by embarking on various user research initiatives + triangulating the results
A comprehensive source of user journey analytics would have been the first port of call to understand where (and potentially why) drop-off occurred. Unfortunately, such information was not easily accessible, due to how the application is not integrated with an extensive suite of product analytics data to analyse

I therefore initiated a multi-sourced user research approach to increase the likelihood that we were able to solve the root cause of the problem by triangulating the results
research sources
💬 User testing
10 x unmoderated task-driven sessions (5 - Desktop, 5 - mobile) via usertesting.com
📺 Screen recordings
Observing circa 2 hours of new user sign-up flows via HotJar, focusing on users who abandoned the profile builder
% Profile completion
I retrieved high level data to understand if there was a trend regarding what sections were not complete at drop off
🕵🏻 Competitor flows
Analysis of benchmarking Tempo’s onboarding flows with comparative products such as Otta, LinkedIn, Feistyjobs etc.
research key observations
Usability: Requesting candidates to input further information within ‘preferences’ in the profile builder created a great amount of friction
  • Supported also by the ‘Design Theory of Proximity’ candidates associated the ‘Preferences’ input to be near the sections of ‘Write a bio’ and ‘Add experience’ based on the checklist component - unfortunately this input is nested within the navigation bar which has poor usability, particularly on mobile
  • It appeared that candidates expected the checklist to be have a linked and scroll to behaviour to the relevant section within the profile builder
  • This learning also indicates why profile completion data suggested that preferences were mostly always incomplete from dropped off candidates, particularly on mobile
Mouse/tap tracking via Hotjar: Users had an expectation for checklist items to do something
The nav bar has poor accessibility, candidates do not correspond the spanner icon to suggest 'job preferences'
An indication of an unmet user goal? : A desire to see jobs to determine the level of effort to commit
  • Candidates frequently interacted with all the ‘See job’ CTAs and any component eluding to the sentiment of seeing jobs throughout the onboarding. If the mandatory profile elements are not yet complete, when pressed candidates are taken back to the profile builder flow
  • The ‘Welcome’ modal also created a lot of curiosity, but its limited interactivity and lack of ability to indicate the types of jobs available on the Tempo platform caused a lot of frustration
  • Competitor flows typically allow users to see upfront the potential value of using the platform by presenting the user with jobs before any effort is required from them
An example of a CTA pointing candidates to 'See jobs'
(Desktop) Welcome modal
hypothesis
Exploring a user goal-orientated hypothesis in a 1 week sprint before tackling usability
Research revealed that enhancing usability could improve the onboarding experience. However, determining the potential value of goal-oriented observation required additional investigation. To address this, I conducted a one-week sprint to assess its potential impact. This was crucial as resolving usability problems constitutes minor tweaks to the user experience, whereas adopting a new user goal perspective could significantly reshape the entire journey.
Currently, candidates can view Tempo's job listings only after completing the onboarding process. 🤔 Hypothesis: Allowing candidates to access job listings before building their profiles might boost motivation, potentially increasing conversion rates
ideation
I led an ideation workshop to explore how might we show jobs to candidates earlier in the onboarding journey
Stakeholders involved - Candidate Lead, Candidate Experience Executives, COO and a Developer. Pros & cons per idea were evaluated against technical feasibility, commercial viability and a consideration of unintended user behaviours an idea may create.
ideation
Low fidelity sketches were produced to review with key stakeholders before proceeding to prototyping
One of product goal of Tempo is to not make finding work feel like a full time job for early career candidates. The flow therefore considered how we can leverage the data that would have been captured to share a snapshot highly relevant job opportunities, with the goal of increasing curiosity to invest further into the product by completing a profile to unlock the ability to see all jobs available on the platform.
My reference point for the 'arc' of my design hypothesis was from article based products such as 'Medium', 'Financial Times' and 'The Times'. The principles of their solution to encourage readers to upgrade (the ideal behaviour) was to show enough of the content for the reader to be curious, before cutting off the content gradually before exposing some copy and the relevant CTA
prototype
I designed a stimulus prototype on Figma to test with users
Scroll down to the bottom of the page to see the 'arc' of the design
user testing sprint
3 x moderated user testing sessions to validate the potential value of the hypothesis
Typically I would engage with 5 participants before arriving at a conclusion, however after participant 3, I had concluded that the cost and time of engaging with 1 or 2 more respondents would not return me with marginally higher learnings, given that the 3 one to one conversations generated insights that were so unanimously aligned
test learnings
User testing revealed that ‘teasing’ candidates with the ability to see more value from the platform creates more frustration than curiosity  🙈
😅 Learnings from the test onboarding experience
  • Candidates felt like we were unfairly withholding jobs from them without justifiable reasoning from their viewpoint. It was perceived to be of no benefit or value to the candidate
  • Showing a ‘sneak peak’ of suitable roles therefore was perceived with a negative sentiment
  • Candidates naturally expect a platform to present them with jobs and recognise that not long after signing up that they would be asked for more information to be presented with relevant content anyways
🤗 Helpful reflections on the current onboarding experience
  • Users have been wanting to ‘see jobs’ in the current onboarding flow because it was suggested as a possibility with various always-on and enabled CTAs
  • Other elements such as the mobile navigation bar as well as the CTAs served as a distraction to candidates and caused confusion
  • The current onboarding is not designed with any success acknowledgement or hand-holding for the user to go down a linear, happy path
test learnings
Defining the goals of the appropriate solution
The sprint to prove the hypothesis helped us to refine the focus areas to improve candidate onboarding
Communicate value
Framing aspects of the onboarding around value provides motivation to get past the friction involved in filling out the form format of the profile builder
Create linearity
Restructuring the flow to create linearity will help reduce the cognitive overload and provide an opportunity for success acknowledgement.
Prioritise mobile
Data indicated that 60% of new signs ups are via mobile, identifying opportunities to optimise further for mobile in the case of creating linearity is highly important
solution
Final outcome: Highlights
Each highlight captures an aspect of the candidate onboarding experience. It calls out which design goal it adheres to, what problem it’s solving and what the solution is.
Create linearity
A logic triggered instruction bar which replaces the previous navigation bar
problem
A confusing navigation UI and lack of success acknowledgement didn’t guide users towards completing the minimum required tasks. The navigation bar also served as a distraction via its highly inaccessible design. Whilst the navigation bar screamed the need for a redesign, as a team we concluded that it should be tackled in a different sprint
solution
We introduced a sticky element to replace the placements of the navigation bar, containing a very clear (and obvious) instruction for users. Once performed, the button within the element transforms from disabled to enabled for the next step, giving candidates feedback of their success
Create linearityPrioritise mobile
Removing the need to navgiate to the preferences page
problem
Candidates were getting lost and frustrated during the onboarding when they could not find the ‘preferences’ section, particularly on mobile
solution
Initially, we directed candidates to the preferences page in the profile builder, requesting data that wasn't utilised in the app. Given mandatory information was already gathered via the chatbot, we streamlined the process by simplifying the completion logic, by making additional preference inputs as optional. This change eliminated unnecessary steps in the user experience.
Create linearityPrioritise mobile
Sticky profile checklist with 'scroll to' behaviour
problem
Candidates expected the profile checklist to have linked and 'scroll to' behaviour to the relevant sections within the profile builder. The experience created a lot of 'rage taps'. In parallel, the feature also appeared to be a frequent reference point for candidates, due how it's an in-line component, we observed friction from candidates continuously scrolling up and down the page to go back to it
solution
We made the profile checklist sticky to the bottom the page. The checklist expands on show, revealing the items, with each item having linked + 'scroll to' behaviour to the body of the page. As a second unintended benefit, it also served as a gamification feature to have the progress bar always on, continuously encouraging candidates to fill up the bar
Communicate value
Refreshing the ‘welcome modals’ with copy that speaks to user goals or pain points
problem
Candidates often gloss over the landing page before proceeding to sign up. When interviewed, candidates weren’t able to speak about how Tempo is different vs. other search platforms. Given Tempo is a relatively irrelevant product in HR Tech, asserting the value of the product becomes important to a user’s motivation
solution
We designed 3 onboarding screens that replaced the existing modal that did not provide much value. Various options of copy tested with our target candidate audience to ensure that the final screens contained copy that resonated with their experiences or goals
Impact
These small refinements increased onboarding conversion by 50% 🤩
... and the percentage continues to grow! At the time of writing (Q4 2022), whilst the figure is small, it is considered a big improvement considering the motivations of candidates to thoughtfully consider job moves is at an all-time low driven by the economic climate. Onboarding continues to be a focus of the product roadmap and further refinements are in progress to increase the figure
retrospect
🙃 Navigating data challenges: triangulating insights for informed design
Despite limited data collation maturity, the absence of visualised user paths posed a challenge in identifying drop-off points. To move forward and minimise risks, we adopted a pragmatic approach, triangulating various data points. This data-driven approach enabled informed decision-making and effective navigation of the design process
🥸 Embracing rejected hypotheses: strengthening questioning skills
Navigating user sessions where hypotheses were rejected proved challenging. However, I swiftly shifted gears, channelling curiosity to make the most of the remaining time. This experience honed my persistent questioning skills, extracting deeper insights and uncovering valuable perspectives