Description

Abstract—LinkedIn is one of the most popular social networks for professionals. Many of us rely on LinkedIn to expand connections and find new opportunities after graduation. In this project, we are going to study one task of LinkedIn – the searching function. First, we will define the problem space, such as what is searching function on LinkedIn and under what environment search function is used. Second, we will outline the user types for this study. Third, we will use 3 different needfinding methods to make a plan, which are Items from the data inventory that will be connected to each needfinding plan. Also, the biases of each needfinding plan will be addressed.
1 PROBLEM SPACE

Figure 1—LinkedIn Search Box
2 USER TYPES
Table 1—User Types for LinkedIn Search Function
User Type User Level of
LinkedIn Working Experience Occupation
Fresh software engineer Novice 1-3 yrs IT
Seasoned trader Expert 4-10 yrs Financial
Senior developer Expert 4-10 yrs IT
3 NEEDFINDING PLAN 1 – PARTICIPANT OBSERVATION
3.1 Needfinding Method – Participant observation
The first needfinding method is Participant observation. In participant observation, the research will be working as a participant observer and observe what are the needs to accomplish the task. In this case, I will be the participant for this exercise and observe different needs to accomplish the task of finding a person on LinkedIn in a different environment.
But currently, in the LinkedIn only Jobs page has a different design (as in figure
2). I probably would expect the search function to search my contact in My Network page and chat history in Messaging page. In the experiment, I will try to search on each page and see if the search result matches my intention.

Figure 2—LinkedIn Search Boxes are the same except in Jobs page
3.2 Data Inventory – Participant observation
Below table summarized data inventories in this participant observation.
Table 2—Data Inventory – Participant observation
Who Where Context Goal Need Tasks Subtasks
Me Desktop My
Network Find co-worker Connect with
colleague Find and Connect to colleague Search by name,
filter by company, send
invitation
Me Mobile Homepage Find Connect Find and Search by
speaker with Connect full name, speaker to speaker send
invitation
3.3 Biases – Participant observation
4 NEEDFINDING PLAN 2 – INTERVIEWS
4.1 Needfinding Method – Interviews
The second needfinding method is Interviews. In this experiment, interview questions will be from the previous participant observation. Through the previous method, I will have some hypothesis from my participation experience. But they could be biased. For example, when I search a person’s name “Oliver”, and would like to apply company name as a filter. On the search result page, if I click the Company button, it will bring me to the result page of companies named “Oliver” instead of using a company name as a filter for people searching results. From the participant observation, I find a need to fix the misleading “Company” button. However, that might not be the same case for other people and it could be a biased need. I will collect interview questions based on my participant observations and ask the user types as defined in section 2.
To make interviews more effective, we need to make sure below 5 points (Joyner,
1. Focus on Who, When, What, Where, Why, How, ask open-ended questions.
2. Be aware of bias, phase questions in a neutral way.
3. Listen, make sure participant do the most of talking.
4. Organize the interview.
5. Practice the questions on friends or families first.
4.2 Data Inventory – Interviews
Below table summarized data inventories in the Interviews method.
Table 3—Data Inventory – Interviews
Who Where Context Goal Need Tasks
Senior Developer Desktop Homepage Teammate leaving
company Connect with
colleague Search function can easily identify teammate’s profile Search by name, find profile and evaluate
Junior PM Desktop Homepage New-
joiner in the Find new-
joiner’s back- Search function can easily identify new-joiner’s profile Search by name, find profile and evaluate
company ground
We can collect the data of interviewee’s feedback about how easy they can accomplish their goal, what are their needs to improve the interface, and also whether the needs from participant observation valid to them. Through this interview process, we can have more insights into the interface.

Figure 3—The “Companies” button is not a filter, but showing company results
4.3 Biases – Interviews
5 NEEDFINDING PLAN 3 – SURVEYS
5.1 Needfinding Method – Surveys
The third needfinding method is Surveys. When we have enough insights about the tasks and needs under different context using qualitative research, we can have more data support by using quantitative research method like surveys. The first two needfinging methods can easily be biased as discussed above. Surveys can help to reduce those biases. Needs from the previous 2 needfinding methods will be rephrased so that the participant can respond in a quantifiable number. For example, a need from participant observation – the confusing button of “Companies” in the search result page should be replaced by filters. In the surveys, a question can be phrased as “On a scale 1 – 5, 1 being the easiest, how much effort for you to find company filter in the people search result”
To make surveys more effective, we need to make sure below 5 points (Joyner,
1. Less is more. Ask minimum questions necessary.
2. Be aware of bias, phase questions in a neutral way.
3. Tie them to the inventory.
4. Test it out. Have co-workers test survey questions before sending them out.
5. Iterate and revise surveys according to the tests.
5.2 Data Inventory – Surveys
Below table summarized data inventories in Surveys method.
Table 4—Data Inventory – Surveys
Who Where Context Goal Need Tasks
Group of
Senior
Developers Desktop Homepage Recall when you were
searching someone on
LinkedIn Successfully Change of found interface Search by name, find profile and evaluate
Group of
Junior
Analysts Desktop Homepage Recall when you were
searching someone on
LinkedIn Successfully Change of found interface Search by name, find profile and evaluate
We can collect quantitative data from surveys that give us a more unbiased view of needs. For each need, we will have a question for that and a series of data points we can calculate the score for that need. For example, if after the survey the need to remove “Companies” button is -5, then it suggests that need is biased to myself and generally it is not a real need for the users.
5.3 Biases
6 NEEDFINDING PLAN CONCLUSION
From the above analysis, we can see that, it’s better to combine several needfinding methods to have more in-depth but less biased needs for our selected user types and problem spaces. Both qualitative research and quantitative research are important methods for us to find the needs of our users.
7 REFERENCES
8 APPENDICES
8.1 LinkedIn Demographics

8.2 Sample Interview Questions
1. What are the common scenarios that you are using LinkedIn search function?
2. What group of people you are using LinkedIn search function to search for?
8.3 Sample Survey Questions
1. On a scale 1 – 5, 1 being the easiest, how easy to find your co-workers by name on LinkedIn My Network page?
2. On a scale 1 – 5, 1 being the easiest, how easy to find conference speakers by name on LinkedIn mobile app?