top of page
Anchor 1

Spekit: Progressive Learning

Track where you are in your learning journey

final-dashboard.png
What is Spekit?

Spekit is a B2B web application and chrome extension where manager can create documents and other resources that can be referenced by anyone in their organization, on any website and through a chrome extension. A primary use case for the app is onboarding new employees.

My role

UX Researcher, UX Designer

Platform

Responsive Website and Chrome Extension

Project Overview

We designed a way for new employees receive content that’s curated specifically for them, by their managers, when they first log in. The goal is to increase retention by showing the value that Spekit provides.

the UX challenge

There was a 37% drop-off between first time login and first time performing an action in our app. In order to increase retention, we needed to research the opportunity to see where we can increase the value.

Chrome search results.png
spek-pop-up-mark.png
topic-partially-completed.png
modal example.png
tracked-topic.png
Frame 281.png
Frame 280.png
Frame 215.png

Learnings

The UX Challenge

Sometimes, good ideas are so big that an MVP would take 6+ months. In those cases, we should consider splitting the MVP into smaller releases and design those releases to provide value to the user while building towards the larger MVP.


As product people, emergencies happen and I’ve learned ways to quickly shift focus to new projects while minimizing loss of work from lower priority projects. I’ve had to juggle multiple projects simultaneously.

Research

known painpoints

We already documented a few first-time user experience pain points from past feedback and collaborating with our customer success team.

 

End-users

  • Current dashboard wasn’t helpful or relevant to new users

  • New users were overwhelmed by the wiki

  • Relevant content wasn’t delivered to new users

  • Some users didn’t know the importance of downloading our chrome extension

user research

We interviewed 10 newly hired internal employees. We observed how they interacted with our existing platform when they first logged in.

findings - existing first time user experience
ex-dash.png
ex-wiki.png
observed typical user flow

Typically, these new hires just clicked around, but didn't find anything relevant enough to read or do.

ex-ftux-user-flow.png
understanding the user journey

It was clear that new users wanted more guidance when they first logged in. We needed some way to highlight content that’s relevant to the user. We began thinking of ideas like:

  • Update “3 steps to multiply your productivity”

  • Step-by-step product onboarding checklist

  • Create a “learning path” where admin can assign Spekit content, PDFs, links to websites, etc.

  • Assign and track 1 topic

user-journey-activation.png

Design

collab with devs and business stakeholders

We created quick sketches and low-fi prototypes to discuss a few ideas with our developers. It helped to get a clearer understanding of the feasibility and timeline of each approach.

We discussed the pros and cons of each idea with our stakeholders.

architecture-brainstorm.png
the BIG idea

As a result, we landed on the idea of a learning path.

Learning path: A progression of topics, curated by a manager, for an end user to read (ie. "Week 1 Sales onboarding" contains topics important to learn in the 1st week of Sales training). Progress would be tracked and be visible to both employee and manager. This could potential grow to include different content types Spekit provides.

the problem with the idea

The idea everyone liked was too big. It needed too many API's that we had to build from scratch and would take 8+ months.

We needed to:

  • create a new product called a “learning path” that could be assigned and tracked

  • create the ability to assign content and track if that content has been read

  • create a way to manage the permissions between different content types at the same time (this alone would take 2 months for our BE developers)

  • provide new analytics for our customer admin

mid-fi prototype and customer feedback

We put together a mid-fi prototype and met with 5 customers to get their feedback. We wanted to understand the most impactful parts and any pain points that they anticipate.

pl-mid-fi.png
findings
  • Viewing permissions can get complicated. Each topics has a unique set of viewing permissions, if a Learning path is a parent to the topic, how will the viewing permissions be defined?

  • The concept of a "learning path" is complicated

  • Managers are hesitant to add a new content type to manage

a small change with big impact

Instead of introducing a new content type, Learning paths, we decided to make Topics assign-able. This eliminated many complexities since we're building off of existing content.

relationship-diagram.png
second (current) iteration
new-dash.png
new-chrome.png
new-analytics.png
dev estimates

My team had a series of meetings where we discussed how we would prioritize the features and divide this project into phases.

  1. Tracking progress (2 sprints)

  2. Assigning topics to a dashboard (2 sprints)

  3. Advanced targeting (tbd)

  4. Set a due date (tbd)

  5. Ability to customize topic order (tbd)

scope-miro.png
validate this plan with customers

With a plan set, we wanted to get feedback from our customers. We updated our mock-ups and prototypes for phase 1 and 2. The goal of this research was to understand if these releases met our customers’ needs.

research.png
findings

We spoke to 11 customers. 10 were excited about the feature. 6 asked to be a part of our beta.

We received a few suggestions that will be considered in future sprints.​

  • This feature would automate our customers’ onboarding checklist

  • Visibility into where individuals are with onboarding is key

  • Users want the ability to change the order that topics appear “logical sequence”

  • Admin want the assignments to be automated for new hires, assign periodically to reinforce knowledge

  • Admin want to be able to customize assignments, assign to teams, users, based on tenure, etc.

  • Some users should have viewing permissions but not be required/ assigned

documentation and design system

In order to ensure this feature was built as designed, we documented all behaviours and micro-interactions that were important to the product. We also documented changes and additions to our design system.

sample of design system documentation
ds-topic-card.png
a shift in priority

Halfway through the development of phase 1, an urgent issue caused our priorities to shift away from this project. Luckily, the majority of what was built enabled marking content as read - a highly requested feature on it’s own. Our team met and decided this would be our initial release.

A benefit of releasing this feature early is for users to start marking content as read. We can start collecting this data for when the main feature is released.

1st release: Mark as Read

Users can now mark any Spek/ content as read, whether it's in the chrome extension, webapp wiki or pop-ups on other websites.

new-mark-as-read.png
new-spek-analytics.png

Next steps

the launch and next steps

As of June 2022, we’ve launched the mark as read feature and the rest of phase one is under development.

Next steps would include:

  • discovery work on advanced targeting

  • research if customers understand the separation of “has view access” vs. “required to read”

  • research how might we add a due date to an assigned topic

bottom of page