Suggested profile summary fully rolled out to production in Jan 2017. Since then it received so many attention from company and users.
1) In company all hands on Dec 2016, suggested summary was mentioned by Jeff (CEO) as best driver for valued members. We had +43.7% lift in quality member profiles within first month.
2) Covered by multiple publisher including Business insider: "On that note, LinkedIn is apparently testing a tool that will auto-generate the profile summary section for some users on mobile phones, based on the content of their profile, as a prompt to get people started"
3) Summary generation was part of LinkedIn intorductory video for brand new Desktop experience.
1) In company all hands on Dec 2016, suggested summary was mentioned by Jeff (CEO) as best driver for valued members. We had +43.7% lift in quality member profiles within first month.
2) Covered by multiple publisher including Business insider: "On that note, LinkedIn is apparently testing a tool that will auto-generate the profile summary section for some users on mobile phones, based on the content of their profile, as a prompt to get people started"
3) Summary generation was part of LinkedIn intorductory video for brand new Desktop experience.
Product Impact
We went with an approach to get the MVP version of the feature out and see if our users discover and accept the suggested summary recommendation with minimal changes. We decided to pick an appropriate template and fill out automatically using information inferred from the member's profile data.
We realized investing into creating client side templates and selection pills to customize attributes to form a summary on the fly would be definitely a V2. But it was great to brainstorm and show the vision of how powerful this feature could be for members who struggle to tell their story on LinkedIn.
We realized investing into creating client side templates and selection pills to customize attributes to form a summary on the fly would be definitely a V2. But it was great to brainstorm and show the vision of how powerful this feature could be for members who struggle to tell their story on LinkedIn.
MVP First Approach
One of the ideas we started to look into was to create certain templates based on our user segmentation and inferred profile data. So we can have separate templates for Students, Job Seekers & Experienced Members based on profile data.
All the templates in the pool are manually reviewed to guarantee quality and grammar consistency.
We looked at all the things people usually mention in a real summary on LinkedIn such as Years of experience, location, current roles, skill expertise and universities for our variable attibutes composing the templates
This was a phase where I was very heavily collaborating with our Data scientist and client side engineers. The weekly meetings and understanding of how can we come to a forward looking solution really sticks out to me.
All the templates in the pool are manually reviewed to guarantee quality and grammar consistency.
We looked at all the things people usually mention in a real summary on LinkedIn such as Years of experience, location, current roles, skill expertise and universities for our variable attibutes composing the templates
This was a phase where I was very heavily collaborating with our Data scientist and client side engineers. The weekly meetings and understanding of how can we come to a forward looking solution really sticks out to me.
Initial Approach to add Customization
We looked at how to best introduce the feature and tease our users who dont have any summary with a glimpse of what the summary could be. That takes them to a takeover screen which prefills the suggested content in the text field. This was the best way to make the flow really easy and frictionless for a lot of our users to add into their profile.
We started on mobile first and tried to make the userflow very intuitive and easy. We adopted the similar approach into our desktop experience.
Part of the design exploration was to see if our techology stack can help us make the auto suggested summaries based on thoer profiles more personale and unique. In order to not have bunch of summaries with look very similar and have a robotic tone to it.
We started on mobile first and tried to make the userflow very intuitive and easy. We adopted the similar approach into our desktop experience.
Part of the design exploration was to see if our techology stack can help us make the auto suggested summaries based on thoer profiles more personale and unique. In order to not have bunch of summaries with look very similar and have a robotic tone to it.
Design & Userflow
By doing text mining of over millions of profiles with well-written and concise member provided summaries, we extracted common patterns and phrases that appear in profile summaries.
We also did a very simple analysis to find the most used phrases in members summary. The average summary length is 82 words. Therefore each summary contains about 4-6 sentences since on average English sentence length is 14-20 words.
Later in the process you will learn how based on these patterns, we created a pool of summary templates that are commonly used among members.
We also did a very simple analysis to find the most used phrases in members summary. The average summary length is 82 words. Therefore each summary contains about 4-6 sentences since on average English sentence length is 14-20 words.
Later in the process you will learn how based on these patterns, we created a pool of summary templates that are commonly used among members.
What we learnt from data
My role & Cross collabroation
Your LinkedIn profile is the best way to tell your professional story. And on your profile, there is no better way to tell your story than filling in the summary section, a top item that recruiters look at.
For many people, writing a summary is a very challenging task. Many members find it a scary and open-ended task and hence they simply avoid it even though they know the value of having one. A lot of our members fill out their profile picture, current positions, past position, location, industry but yet they do not have summary. Clearly there is a pain in writing a summary.
1) Our main goal is to encourage our members to have profile summary by providing suggested summaries for those who have rich profile information but no summary.
2) We can also provide suggestions to those members whose summary can be improved in some ways.
For many people, writing a summary is a very challenging task. Many members find it a scary and open-ended task and hence they simply avoid it even though they know the value of having one. A lot of our members fill out their profile picture, current positions, past position, location, industry but yet they do not have summary. Clearly there is a pain in writing a summary.
1) Our main goal is to encourage our members to have profile summary by providing suggested summaries for those who have rich profile information but no summary.
2) We can also provide suggestions to those members whose summary can be improved in some ways.
Project Goals
“Profile Summary” is one of the most important piece of information on a member’s LinkedIn profile. It is mainly the first section that viewers read after landing on a profile page. Sometimes recruiters filter the candidates by looking at their summary only.
It goes without a doubt that having a decent cover letter or bio increases a candidate chance to get the job offer. By the same reasoning, having a summary increases the member chance to get reached out from recruiter or receive and inMail/connection request from other members.
It goes without a doubt that having a decent cover letter or bio increases a candidate chance to get the job offer. By the same reasoning, having a summary increases the member chance to get reached out from recruiter or receive and inMail/connection request from other members.
Overview
We want to make writing Profile Summary easy and simple on LinkedIn