In UX we keep getting asked to bring a frictionless future... to nowhere

Thanks to the many people who have reached out with leads on available senior-y leveled data/ux positions or just connections to other humans! I'm working through things and starting to reach out to folks. Also, thanks to everyone who just reached out with kind words. It means a lot.

Meanwhile, I'm going to continue writing as always, though current events might bias my interests and focus in certain directions. I'm not giving up on a 6-year streak. Also, thanks to subscribers, this newsletter's operational costs are covered, so assuming I can earn enough to support my family, we can keep going forever.

To the surprise of nobody, it's been a very busy, and stressful, one week. Honestly, the only thing counteracting the overall lethargy and depressive mood of being hit with a layoff is my internal neurosis of needing to get my stuff in order and safe before I can sit down and relax. Luckily, I've been through layoffs many times before in my career so I've got good coping mechanisms. I'm also getting enough sleep.

So today I want to talk about one of the greatest "Standard Enemies" within UX – Friction.

As a quant UX researcher, and honestly as ANY kind of UX researcher that is supporting engineering and product development teams, one of the most common types of work we are requested to do is "find and reduce friction in a process".

A classic example is in a web store's checkout flow. If the checkout process is extremely burdensome to the user, with long forms, lots of confusing detail, missing interface elements, then users are likely to get frustrated by the whole process and abandon their attempt to purchase. So the general business theory is, it should be as simple as possible for users to give us money. Because you'll overall get more money in the process.

Now, there is an extreme example of this that is obviously true. If you were about to checkout to buy $15 worth of snacks and I forced you to do a 50-question survey before you were allowed to complete your order, you'd rightfully get pissed and stop. That survey is ridiculous unnecessary friction to the checkout process and it doesn't take a usability genius to point out that it is a problem and advocate for removing it. A lot of UX work comes from identifying the less extreme examples of the same pattern, because no one likes doing unnecessary work.

But humans are fascinating creatures. There comes a point "optimizing" a checkout flow where you start creating more problems in the name of lowering friction. In this example, it'd be those "1-click to buy!" buttons that exist in various stores. You hear stories of how parents with little kids who get their hands on Kindles rack up all sorts of purchases with the feature and there's plenty of threads on the internet asking for advice on how to disable the feature or at least get refunds.

What has happened is that beyond some critical point, you stop lowering friction and instead start shifting friction somewhere else. Pavel Samsonov, a UXer I respect quite a lot, writes about this in various places. "Making things easier" is not by itself a goal that makes better products. Users don't buy things from your store for the sake of checking out. They don't recommend your products because it didn't hurt to pay you. They're buying stuff to serve whatever need they have. And when you make things too easy, too frictionless, you create brand new never-before-seen headaches like "my kid just bought 100 comic books on my device and I can't return them".

The shuffling of friction

So the reason why I've got friction on my mind right now is because I'm job hunting! The last time I was actively on the job market was around 2019. I have very strong memories of searching for work in the 2008 crash where I did most of my job searching on places Monster and Craigslist of all places. Then the web 2.0 explosion led to all sorts of innovation in the job hunting space with people trying all sorts of interesting business models and interfaces.

But now, the reality of 2025 job hunting is that much of the innovation had really stopped in the late 2010s. There's only so many variations of "match job applicants with job openings in a big algorithmic market-like situation" that can be done. Instead, innovation has long devolved into "making it easier" for applicants to apply to things. After all, if everything about job hunting is a number's came, then the best way for a job listing company to increase their metrics on "jobs filled" is to ram as many applicants as possible through.

The end effect of modern job system interfaces reducing friction has resulted in massive explosion of applicants to every position you can imagine. It's entirely possible now for me to just indiscriminately mass apply to positions by blasting my resume out and making a handful of clicks. The cost in terms of effort to me is extremely low, while the cost of reading the applications becomes sky high. And this is how we hear how hiring managers are absolutely FLOODED with applications now, many of which are unqualified garbage. The LLM boom right now has made this even more painful since now every application might be a partial or complete fabrication. At this point, many people have declared the job market to be fundamentally broken. I honestly do not see evidence to the contrary.

And so, the friction was moved from applicants to HR departments. And in response, HR has implemented automated filtering techniques to cut down on the noise. To their mind, the chance of missing some highly qualified applicants due to an imperfect filter is still preferable to being completely overwhelmed. I don't blame them, even though this directly affects my chances at finding a position right now. The best candidate out of 200 is probably close enough to the best candidate out of 10,000. It is an arms race with no regulating factor. It's terrible and shouldn't be this way.

But because it's an emergent interaction between self interested job seekers and job purveyors, there's no good way to get it to stop.

The challenge of zooming out

The above problem of unintended negative side effects has always been a bit of a fascination with me. I've looked at this problem before from the perspective of "if all our GPS navigation systems utilize traffic data to pick routes for you, and everyone uses such a system, traffic becomes a normalized spread phenomenon where no one can get an edge any more". There's also been articles on the internet about how map navigation has started leading more and more traffic down side roads that aren't designed to handle the load, causing new, potentially unsafe, traffic issues.

There have been many times in my career where a PM desperate to hit results have asked me to help find more friction to reduce in some product flow, hoping that the next insight will be the one that brings the hockey stick growth they promised their managers. There have honestly been times in such meetings I had to keep myself from shouting that there is no juice left to squeeze, please abandon this work and do something more impactful. We are not ethically allowed to force users to buy a product when their eyes simply land on the logo.

Pushing back on this appropriately (read: without screaming) is quite a bit of work since there's obviously not a lot of alignment in goals. They want the product they're responsible for to do well, while we don't want to spend a minute more on what is obviously a lost cause request.

But I'll admit up front that getting to the point where I realize that something is a lost cause is slow. The first few interactions and projects are usually in the "it's worth a try" category – there's still obvious points of unnecessary friction stopping users from doing their work. There's still cruft to remove and streamline. But over time you see how little effect the improvements have and realize that "making stuff easier" is not the root problem. Whatever adoption issues exist it lies outside, either in feature discovery, or it doesn't fit user needs and they're pointedly telling us with their clicks.

What keeps me up at night is this open problem of how do we learn about these unintended side effects? How long did it take for Amazon to learn that 1-click was causing issues for parents? How long did it take for people to realize that the cause of weird new traffic patterns was the universal adoption of navigation aids? How many people had to have their lives affected in some way for complaints to aggregate into our complex metrics platforms to trigger a response? To what extent can we zoom out and predict at least some of these effects enough to be on the lookout for them?

I don't have a definitive answer, but I do know that it involves qualitative research work. These edge cases usually start off small and as anecdotes. They'll slowly snowball as things worsen, but if you dig deep into the data you'll see issues below the radar start almost very soon. It only comes to our attention when enough pain and escalations happen to surface it. I just wish we could all have avoided the pain to get to that point.

This is far from a solved problem. In fact, I know a fair number of people who would declare that it's "someone else's problem" due to how incentive structures are often twisted in product work. I don't respect those people very much. Unfortunately, as the people who are responsible for bringing the voice and needs of the user to the design table, it's our problem as people who care about user experience.

Standing offer: If you created something and would like me to review or share it w/ the data community — just email me by replying to the newsletter emails.

Guest posts: If you’re interested in writing something, a data-related post to either show off work, share an experience, or want help coming up with a topic, please contact me. You don’t need any special credentials or credibility to do so.

"Data People Writing Stuff" webring: Welcomes anyone with a personal site/blog/newsletter/book/etc that is relevant to the data community.

About this newsletter

I’m Randy Au, Quantitative UX researcher, former data analyst, and general-purpose data and tech nerd. Counting Stuff is a weekly newsletter about the less-than-sexy aspects of data science, UX research and tech. With some excursions into other fun topics.

All photos/drawings used are taken/created by Randy unless otherwise credited.

• randyau.com — homepage, contact info, etc.

Supporting the newsletter

All Tuesday posts to Counting Stuff are always free. The newsletter is self hosted. Support from subscribers is what makes everything possible. If you love the content, consider doing any of the following ways to support the newsletter:

• Consider a paid subscription – the self-hosted server/email infra is 100% funded via subscriptions, get access to the subscriber's area in the top nav of the site too
• Send a one time tip (feel free to change the amount)
• Share posts you like with other people!
• Join the Approaching Significance Discord — where data folk hang out and can talk a bit about data, and a bit about everything else. Randy moderates the discord. We keep a chill vibe.
• Get merch! If shirts and stickers are more your style — There’s a survivorship bias shirt!