Reflecting 2018 as a UX researcher

George Zhang
4 min readDec 21, 2018

During my last working week of 2018, I spent a substantial amount of time in reflecting my experience as a user experience (UX) research leader at Uber. I usually did it during commutes either in an Uber car or on a BART. Once I had something worthwhile in mind, I wrote a FB post to gather comments from my friends there.

2018 is awesome, wonderful and insightful in the rear view mirror as it’s passing from sight! I learned a ton, and would like to share several points below. Again, all of them are my personal opinions that welcome your challenges and comments. Yes, a food for thought!

1. Serve the needs that underpin what users want

In order to change how people behave in a new ecosystem, serve what they need that underpins what they want.

First of all, don’t be misled by the latter which quite often twists what people truly need. The twisting factors could be anything ranging from peer pressures, social acceptance to a mere attention bottleneck. Discovering and then serving the unaware, untold, or unmet need should always be our target as a UX researcher.

This principle can be a barometer to judge a researcher’s handle over their craft.

2. Need is an internal imbalance

A need is often an internal status of imbalance, like hunger, thirst, or mobility, that drives people that realized it to desire a change that can bring back the balance which in their view is a better self.

In industry, operational or technical imbalances are often mistaken as user needs, and it’s the researcher’s job to find true imbalances in the target users to correct the mistakes.

How do you define a person’s need?

It’s hard to define a need without context. What I found useful is to look at the moments in the journey context to achieve a clearly identified goal. For example, a searcher needs query formulation assistant when files a query, a rider needs to know the driver and car that will pick her/him up, etc.

Actually many easy-to-implement methods can help us find needs, such as the classic task analysis and the newcomer “job to be done”.

3. MSL and JND matter a ton

It’s useful to understand and measure minimum sufficient level (MSL) and just noticeable difference (JND) in the target population.

The MSL refers to the level of a specific attribute that satisfies users to the extent they choose to stay on the platform. Yes, it’s a sustainability metric which is very important for marketplaces. MSL can be measured and tracked by the retention and engagement metrics of the HEART framework which was advocated by Kerry Rodden.

The psychophysical JND refers to the level that 50% of users realize that a specific designed solution and an alternative are different on a specific attribute.

Know the two, and make sure the new solution delivers an experience way above and beyond.

4. User delight is difficult to design at scale

What user delight is in design is something easy to argue about but difficult to agree upon.

My understanding is that delight is a moment that stimulates both psychological arousal (surprise) and psychological valence (pleasure) above and beyond what users had anticipated. To put it straight, delight is a surprising pleasure or a pleasing surprise.

Under-indexing either one ruins the designed delight, and I saw more cases that under index the surprise factor when a delightful design gets over used over time or fixated to a place of a user journey. Consequently users get a fatigue and can easily anticipate exactly what will happen.

Yes, delight in my mind is difficult to design and deploy at scale, but so easy to ruin.

5. Define and manage early adopters for an MVP

MVP (minimum viable product) is meant to be used by, ideally only be used by, the early adopters of a new product or a new service, to gather timely user feedback for quick iterations. Deploying an MVP to a small sample indiscriminately selected from all walks of life is a good recipe to fail miserably.

A UX researcher needs to not only help define the target user population, but also help define and manage the early adopters in the target users. Yes, a tough but interesting job.

6. Things go worse before getting better

It can be a mantra for my new year, and it’s definitely true in the experiments I saw in 2018. Things often go worse before getting better, even through the true experiments upon a well designed and implemented product change.

What a researcher shall do often is to optimistically communicate again and again why we have to build it to serve the end users, and push it forward until lasting positive impact is established. Yes, a good change sometimes takes both time and persistence.


My great thanks go to Molly Stevens, Tomer Sharon, Darya Pilram, Naman Mathur, Zhou Bailiang, Paula Desmond, Ramona Del Mar, Paul Jaye, Larry Wei, Jessica Tsai, Ying Liu, Matthew Sola, Ginger Kimler, etc. who commented and contributed on my FB stream before I composed this article. I also incorporated their feedback in this article.

I also deeply appreciate the many more likes my FB posts received from all other friends. Honestly without their timely support, I can’t have this article to share on Medium. Thank you all, my friends on FB!

Lastly, I need to thank my daughter Sharon Zhang, a fifth grader, for proof-reading this article on Saturday.



George Zhang

Global Head of Product Design, Brightly a Siemens Company. Formerly Google, Uber, Intel, Course Hero. Received Ph.D in I/O psychology.