Redesigning LinkedIn's Messaging Interface

LinkedIn Messaging screenshot.png

It was our very first day in the PMDojo Accelerator program and while we were all pitching our own problem ideas to form teams, I heard one of the cohort participants ask us “Do you know what I don’t like about LinkedIn’s Messaging? The fact that you need to scroll through all the messages, just to find one specific message that you are looking for!”  A lightbulb went off in my head, and I remembered the countless times I have scrolled through conversations trying to find messages. Then and there, I knew that this was exactly the problem space that I wanted to explore and solve as a Product manager. So I joined this team and together our goal  was to work on creating a Minimum Viable Prototype (MVP) for the next 8 weeks.

Context of the Problem

As we started working on the problem, my team and I articulated the problem statement as

“LinkedIn users are expressing annoyance because it is hard for them to locate a particular message within a conversation.” 

Assumptions

We used Philipp Krehl’s Uncertainty Rubric to assess all the areas that we were uncertain about and identified the below prioritized list of assumptions.

  • There is an overwhelming need to fix this current problem. 

  • Existing products in the market have features that could be leveraged to solve this problem.

  • Users might be willing to pay for this potential solution.

  • The potential solution will improve productivity and increase user engagement on LinkedIn’s messaging platform.

Risks 

We also identified risks that we could run into such as not focusing on the right problem, targeting the wrong audience and that customers dislike the potential solution. The next steps were to validate our key hypothesis and assumptions as early as possible so that we do not spend our time and energy on a pseudo problem.

Identify Target Customers

Now the burning questions were-

  • Who are our target users? 

  • How do we reach out to them? Would surveys be more effective or user interviews? Or should we just create a poll? 

  • What do they like and dislike about LinkedIn’s messaging interface? 

  • Who are LinkedIn’s competitors – both direct and indirect?

Target Market

LinkedIn is no longer just a simple network for professionals to connect, hire or get hired. LinkedIn has evolved, diversified and classified its business solutions in the following four categories: Hire, Market, Sell and Learn.  For the purposes of our case study, we decided to only focus on the ‘Hire’ category, and the target users that have the ‘free Basic’ LinkedIn Account, and LinkedIn ‘Premium’ Account (career).

Target Users

For our target market, we identified the following target users: 

  • Thought Leaders – Experts in a field and Mentors that influence others.

  • Job seekers – Recent graduates that are searching for a job and Professionals looking to change their career that want to network with Recruiters, Mentors and like minded individuals.

  • Recruiters – Recruiters looking to attract and hire the right candidates.

Since we did not have enough data to draw insights for thought leaders, we decided to narrow the target users down to ‘Job Seekers’ and ‘Recruiters’.

User Personas

Let me know introduce you to our user personas: 

Meet Sheetal, a 33-year-old Architect looking to pivot into Product Management

Sheetal Job seeker.png

And Phil, a 42-year-old Senior Recruiter at TechSmart Co.

Phil Recruiter.png

Market Validation

Given that we had narrowed down the target market, we still did not know our users' pain points and if the problem we were trying to solve was indeed a problem?!?  The tools that we used to validate our market were user interviews, surveys research and competitive analysis. 

User Interviews

We divvied up the task of interviewing our users. This was an excellent opportunity for me to learn how to conduct user interviews. Being an advocate for empathy, I was able to put myself in the user’s shoes and understand their pain points. Grateful to the 4 job seekers and 1 recruiter who offered their time and were willing to speak with me over phone/via video call. 

Here are some of the quotes from the User Interviews

“LinkedIn Messaging is super easy to use, however I have to scroll up and down to read messages and I don’t have time for that” – Priyanka, Recruiter.

“I feel safe using LinkedIn messaging as I can contact an unknown person without sharing any personal information and have seamless visibility to their profile” - Mebi, Job seeker.

“I love that I can send quick short professional messages, but I switch to other platforms as recruiters respond slower on LinkedIn” - Lisa, Job seeker.

“I want to be able to prioritize my conversations, so that I can easily find important messages” - Jithesh, Job seeker.

“I normally stick to LinkedIn messaging itself, and I like the seen feature. I do wish though that I had the ability to save my messages” – Rincy, Job seeker.

From the user interviews it was evident that while our customers thought that LinkedIn messaging was intuitive and easy to use, they were not entirely satisfied with the messaging experience on LinkedIn and were switching to other alternate messaging platforms. 

Survey Analysis

To validate our assumptions with quantitative data, we designed our survey primarily with percentage-based questions. We sent our survey to professionals within our career circle and posted it in product management groups, and we were elated to receive 40 responses. A big thank you to everyone that took the time to answer our survey!!!

Through our survey, we wanted to understand what type of accounts our target users hold? How much time do they spend messaging when they are on LinkedIn? What are the reasons for them to chose another messaging platform over LinkedIn (if they do)? 

Below is a snapshot of our survey results:

What type of account do you currently have on LinkedIn?

Survey1.png

While on LinkedIn, what % of time do you spend messaging?

Survey2.png

If you do switch to another platform, what are the percentages for the reasons of moving the conversation from LinkedIn messaging?

Survey3.png

Which messaging platform do you switch to?

Survey4.png

Competitive Analysis

Since LinkedIn's competitive landscape is huge, we decided to divide and conquer the research. I found that LinkedIn is significantly ahead of its 2 direct competitors - XING and Viadeo. 

It was interesting to see the popularity of key word "LinkedIn" on Google Trends in different countries worldwide. 

linkedIn popularity.png

Even though Xing and Viadeo are LinkedIn's direct competitors, for the purposes of our case study, since we are focusing specifically on the messaging aspect of LinkedIn, it made more sense for us as a team to dive deeper into the indirect competitors such as Email, Text messaging, WhatsApp and analyze what they already offer our target users.

Collective Inference and Reframing the Problem

From the surveys, user interviews and competitive analysis, we validated our assumptions and drew insights into our target users’ pain points and realized that LinkedIn has a lot of opportunities that they can tap into. Hence, we decided to reframe our problem statement and craft LinkedIn Messaging's Value Proposition Statement.

“LinkedIn users are expressing annoyance because of a lack of features in the messaging interface and are thus switching to other platforms.” 

Value Proposition Statement

Our goal is to help our target users (job seekers and recruiters)

Who want to effortlessly use LinkedIn to not only network but also communicate 

By providing features that make the messaging experience both seamless and efficient 

Unlike our competitors

Road to Strategy

Utilizing our research, we were able to identify and list down the user needs. Then we used the Kano model to categorize these needs into three buckets- ‘Must Have Needs’ (Basic Table stakes), ‘Performance Needs’ (if solved, will increase customer satisfaction significantly) and ‘Delighter needs’ (which will induce excitement in our customers).These three needs collectively constitutes the Customer Experience of the product. 

Key Metrics

We identified the following metrics to be the key indicators to define and measure success:​ 

To get a complete snapshot of the metrics, they will have to be measured pre launch to get the AS-IS numbers. Then 2 weeks post launch,  measure this daily for 3 months. After which, they can be measured every month for 1 year (based on the performance of the metrics, this time frame can be reevaluated).   

Prioritize and Evaluate Tradeoffs

By this time, we had put our heads together and come up with several ideas to solve our user's pain points. Then we started writing use cases and broke them down into user stories . The next step was to analyze the Impact and Effort of these user stories and plot them on a prioritization matrix. 

Impact Effort Matrix.png

*Scale = 1 to 5, where 1 is the lowest and 5 is the highest

The whole idea of the prioritization matrix is to evaluate the Effort needed to build an idea and the Impact it would provide our users. The user stories with Low Impact and Low Effort are the 'Fill Ins' and 'Thankless Tasks', none of our user stories fell into those categories. The 'Quick Wins' are the ones with High Impact and Low Effort. And, the 'Major Projects' are ones with High Impact and High Effort. At this point we had to evaluate tradeoffs. This does not mean that the ideas were not good, and they could still be used as a back up plan. However, taking into consideration that the Accelerator program is only 8 weeks long and that we were already half way through the program, for our Minimum Viable Prototype, we decided to focus on three pain points that would solve a 'Must Have Need', 'Performance Need' and 'Delighter Need' as demonstrated in the 'Problem Roadmap' below. 

Problem Roadmap3.png

Glimpse of our Product Backlog

Below is a screenshot of our product backlog in Jira with the user stories that we will be working on for our MVP:

JIRA product backlog1.png

Wireframes of the User journeys

Before jumping right into building clickable prototypes for the solution candidates, my team and I wanted to take a step back and create low fidelity prototypes. Below are the wireframes that I created  to demonstrate the user needs/ pain points for our user personas - Phil (the senior recruiter) and Sheetal (the job seeker).

Use case 1

Phil is overwhelmed with the sheer number of messages that he receives daily – most of them being similar queries but he must ignore the messages due to limited time. He is concerned that this could cost him to miss out on the right candidates.

Phil Use Case.png

Use case 2

Sheetal is constantly reaching out to recruiters and does not want to lose track of important messages and conversations.

This morning, she is late for an interview that was supposed to start at 10:00 am because she was unable to find the meeting invite details that the recruiter referenced in their message. Sheetal is upset and she goes on social media to complain about LinkedIn’s messaging interface.

Sheetal Use Case a.png

A few days later, Sheetal notices an Ad about new LinkedIn Features for messaging and she is super excited to try these features out.

Sheetal Use Case b.png

After LinkedIn rolls out these features, Sheetal is now a happy camper.

Sheetal Use Case c.png

Wireframe Validation

As a team, we validated my wireframes with 3 job seekers and the consensus was that they did not like the icons that I used for saving messages (bookmark symbol) and prioritizing important conversations (star symbol). Instead they asked if we could use the “Save” symbol that LinkedIn currently uses for saving posts and a “Pin” symbol to prioritize important conversations.

Solution Candidates 

We already had a fair idea of how to solve the ‘Must Have’ need and ‘Performance’ need. However, coming up with a ‘Delighter’ feature was the toughest part of this project. It took days of brainstorming and having to resolve tension within the team to finally agree upon the most optimal solution.

As a learning exercise, each one of us in the team created clickable prototypes for these solution candidates. Due to time constraints we decided to only create prototypes of the iPhone LinkedIn Mobile app. I had a lot of fun exercising the creative side of my brain while designing my prototypes on Balsamiq and InVision.

Ability to SAVE Messages

Check out my prototype here. For better clarity of the video, please watch it on my YouTube channel

Use case - As a job seeker, I want to be able to save messages within a conversation, so that I can easily locate messages in the future.

Snapshot of a User Story in Jira

We wrote all our user stories in Jira. Attached below is a screenshot of one of the user stories just to demonstrate how descriptive user stories can get. What I learnt through this process is that a user story should be able to exist on its own and still make sense. It should be granular enough to be able to complete in a single iteration. Defining the Acceptance Criteria is important for the story to be marked as done. Last but not the least, attaching screenshots of the prototype will help clarify requirements.

Jira User Story SAVE.png

PIN Feature

Use case - As a recruiter/job seeker, I want to be able to pin conversations so that I can prioritize important over unimportant conversations in the messaging interface.

Hybrid Chat

One of the pain points that job seekers had voiced during our user interviews was the radio silence that they experience after reaching out to a recruiter. While we were exploring this problem and envisioning the solution candidate, I did some research and found interesting statistics that validated the need for a ‘human + bot’ experience. 

  • Per a CareerBuilder survey, 67% of job seekers have a positive impression of a company if they receive consistent updates throughout the application process.

  • According to Gartner’s research for Talent Acquisition, by 2022, 35% of organizations will turn the job application process into a simple conversation using Natural Language Processing in their recruiting process. 

  • In a survey conducted by Allegis, 58% of candidates were comfortable interacting with AI and recruitment chatbots in the early stages of the application process.

For our MVP, we plan on introducing the Hybrid Chat feature where recruiters can configure Auto Responses to common queries, and they will also have the ability to enable or disable this functionality. In future releases, it would be a great idea to enhance the Hybrid Chat functionality to schedule interviews, collect resumes and store them in a folder in the LinkedIn messaging interface for the recruiter to review later.

Insights from User Testing

After creating our own prototypes for the Hybrid Chat, we decided to conduct user testing on the prototype. For my prototype, I reached out to 2 recruiters. 

“We don’t intend to ignore job seekers; it is just that most of the time we are overwhelmed with the number of job seekers that reach out to us. Such a tool will help with pre-screening some of the basic questions. The only thing is it must be easy to configure since many of us are not tech-savvy” – Priyanka, Recruiter.

Here is the feedback that we received from all the user testing

  • Make it as Human and Personal as possible. 

Our team thought that one way for Recruiters to do this is while configuring the Auto Response messages, they could ensure that              they are addressing the job seeker by their name.

  • There must be a way to differentiate between the Recruiter’s response and the BOT auto response.

 Therefore, our team decided to introduce the text ‘AR’ next to the message – to indicate it is an Auto Response. We also decided to         include a tooltip next to the text  ‘AR’ to help our users to understand what it means.

  • Need ways to customize responses based on the job seekers message.

 Hence, I modified my prototype and created an alternate path to demonstrate this.

Please check out my fun and interactive Hybrid Chat video prototype here 

Use case - As a recruiter, I want the ability to auto respond to common queries so that I do not keep my connections waiting.

User Story Mapping

We also did a fun white boarding session using the collaborative tool Miro where we mapped out our User Stories and picked only those stories that we will be working on for our MVP. 

Before Mapping

After Mapping

miro.png

Release Planning

To ensure that outcomes are achieved, we created a tactical actionable plan to track the features that we have planned for the MVP. There will be a total of 4 Sprints each 2 weeks long.

Product Roadmap.jpg

Launch Readiness Checklist

Summary and Conclusion

Working on this case study was an excellent opportunity for me to hone in on my Product Manager skills by redesigning the messaging interface on LinkedIn. The goal is to improve the messaging experience for job seekers and recruiters by making the messaging interface more inviting and therefore increasing productivity and user engagement on LinkedIn. To achieve this, our Minimum Viable Prototype will consist of the following features – Save specific messages within a conversation, Prioritize important conversations and a Hybrid chat functionality to auto respond to connections.

If you have any other ideas on how to improve LinkedIn's Messaging experience, please feel free to reach out. I would love to connect with you and chat about it!

Key Resources