Syndicate content LukeW | Digital Product Design + Strategy
Expert articles about user experience, mobile, Web applications, usability, interaction design and visual design.
Updated: 21 hours 37 min ago

Ask LukeW: PDF Parsing with Vision Models

Tue, 06/25/2024 - 2:00pm

Over the years, I've given more than 300 presentations on design. Most of these have been accompanied by a slide deck to illustrate my points and guide the narrative. But making the content in these decks work well with the Ask Luke conversational interface on this site has been challenging. So now I'm trying a new approach with AI vision models.

To avoid application specific formats (Keynote, PowerPoint), I've long been making my presentation slides available for download as PDF documents. These files usually consist of 100+ pages and often don't include a lot of text, leaning instead on visuals and charts to communicate information. To illustrate, here's of few of these slides from my Mind the Gap talk.

In an earlier article on how we built the Ask Luke conversational interface, I outlined the issues with extracting useful information from these documents. I wanted the content in these PDFs to be available when answering people's design questions in addition to the blog articles, videos and audio interviews that we were already using.

But even when we got text extraction from PDFs working well, running the process on any given PDF document would create many content embeddings of poor quality (like the one below). These content chunks would then end up influencing the answers we generated in less than helpful ways.

To prevent these from clogging up our limited context (how much content we can work with to create an answer) with useless results, we set up processes to remove low quality content chunks. While that improved things, the content in these presentations was no longer accessible to people asking questions on Ask Luke.

So we tried a different approach. Instead of extracting text from each page of a PDF presentation, we ran it through an AI vision model to create a detailed description of the content on the page. In the example below, the previous text extraction method (on the left) gets the content from the slide. The new vision model approach (on the right) though, does a much better job creating useful content for answering questions.

Here's another example illustrating the difference between the PDF text extraction method used before and the vision AI model currently in use. This time instead of a chart, we're generating a useful description of a diagram.

This change is now rolled out across all the PDFs the Ask Luke conversational interface can reference to answer design questions. Gone are useless content chunks and there's a lot more useful content immediately available.

Thanks to Yangguang Li for the dev help on this change.

Ask LukeW: Text Generation Differences

Thu, 06/20/2024 - 2:00pm

As the number of highly capable large language models (LLMs) released continues to quickly increase, I added the ability to test new models when they become available in the Ask Luke conversational interface on this site.

For context there's a number of places in the Ask Luke pipeline that make use of AI models to transform, clean, embed, retrieve, generate content and more. I put together a short video that explains how this pipeline is constructed and why if you're interested.

Specifically for the content generation step, once the right content is found, ranked, and assembled into a set of instructions, I can select which large language model to send these instructions to. Every model gets the same instructions unless they can support a larger context window. In which case they might get more ranked results than a model with a smaller context size.

Despite the consistent instructions, switching LLMs can have a very big impact on answer generation. I'll leave you to guess which of these two answers is powered by OpenAI's GPT-4 and which one comes from Antrhopic's new (this week) Claude 3.5 Sonnet.

Some of you might astutely point out that the instruction set could be altered in specific ways when changing models. Recently, we've found the most advanced LLMs to be more interchangeable than before. But there's still differences in how they generate content as you can clearly see in the example above. Which one is best though... could soon be a matter of personal preference.

Thanks to Yangguang Li and Sam for the dev help on this feature.

Ask LukeW: Dynamic Preview Cards

Thu, 05/30/2024 - 2:00pm

After adding the Ask Luke feature to my site last year, I began sharing interesting questions people asked and their answers. But doing so manually meant creating an image in Photoshop and attaching it to posts on Twitter, LinkedIn, etc. Now with dynamic Open Graph previews, these preview cards get created on the fly- pretty sweet.

Ask Luke is an AI-powered conversational interface that uses the thousands of articles, videos, audio files, and PDFs I've created over the years to answer people's questions about digital product design. Every time the system answers a question, it does so dynamically. So technically, each answer is unique.

To make each question and answer pair sharable, the first step was to enable creating a unique link to it. The second was to use Vercel's image generation library to create a preview card each time someone makes a link.

The dynamic preview card for each question and answer pair includes as much of the question we can in addition to a bit of the response. It also adapts to varying question and answer lengths since it is generated dynamically.

When shared on Twitter, LinkedIn, Apple Messages, Slack, and any other application that supports Open Graph previews, an image with the question and answer is displayed providing a sense of what the link leads to.

Thanks to Yangguang Li, Thanh Tran, and Sam for the tips and help with this.

Bolting on AI Features

Tue, 05/28/2024 - 2:00pm

As more companies embrace new AI-enabled capabilities, a commonly held position is that established players will "win" by integrating AI features into their existing platforms and products. But the more established these companies are... the more competing interests they face when doing so.

Consider Microsoft's Web browser Edge and its start-up experience. Edge is now your AI-powered browser. But it's also your way to browse, shop, find, create, game, protect, learn, pin, personalize, sign-in, import, sync on the go, and discover. In other words, any new AI feature faces stiff competition from all the other existing features that are still vying for people's attention and use. (Lots of internal team objectives to hit at Microsoft)

Sure, the AI features are mentioned first but they're likely tuned out and skipped over like all the other browser features being promoted during setup. After years of being asked to adopt sign-in, shopping, syncing, personalization and more, people have learned to ignore and dismiss marketing messages especially when they come as fast and furious as they do on Windows.

And it's not just setup. Once you start using Edge, the right side-panel is loaded with icons. Of course, the most brightly colored icon is the AI feature but it's right there alongside search, shopping, tools, games, Microsoft 365, Outlook, Drop, Browser essentials, Collections, and more. Once again the effect is to tune it all out. Too much, too often. And the new AI features fare the same as all the existing features.

Looking beyond Edge, the same issues persist across Windows. Yes, there's a new AI feature icon but it's competing with Windows Start, Microsoft Start, and god knows what else (after a while I was afraid to click on anything else).

Contrast this situation with new products and companies that start with AI capabilities at their core. They don't have a laundry list of pre-AI features competing for attention. They are not beholden to the revenue streams and teams behind those features. They can build from the ground up and use AI-based capabilities to build the core of their offering leading to new paradigms and value adds.

Of course, new entrants don't have massive user bases to leverage. But often a large existing user base is a disadvantage, because adoption of new features isn't earned. Bolt on an AI feature to existing user base and some subset will try it or use it. The numbers "look good" but even turkeys fly in hurricanes.

Building from the ground up means you have to earn each user by providing value not just promotions. But it also means you're creating something valuable if people decide to come and especially if they stay. When you're integrating features, it's often harder to tell.

Mind the Gap

Tue, 05/07/2024 - 2:00pm

Despite good intentions, lots of user-centered design isn’t actually user-centered. Learn what drives these gaps and how your organization can align business and customer needs to deliver the kind of user experiences we all want to have online. With data informed insights, “live” redesigns, and more Luke will give you the tools and information you need to close the gap between customers and companies.


Hello everyone, today I'll talk through the mobile opportunity and how we're doing delivering actually user-centric design and products to the world.

I often like to say mobile is a planet-scale opportunity. But what do I mean by planet-scale. Well to start, there's 7.7 billion people on our planet. Of those, not everyone can use the software and services we make. Some are too young, some are too old, so we need to slice this number down a bit. It's not perfect, but 14 plus is one way of doing that.

Of the 7.7 billion people on the planet, about 5 billion have a mobile plan. Some with data, others not, but they're subscribed to something on their device. And last but not least, there's about 4 billion active smartphones.

While these smartphones vary in their capabilities, they're effectively as powerful as personal, even supercomputers of the past. They're pocket-sized and connected to the global internet pretty much all the time. With 4 billion of them active today, that's quite the achievement and quite the opportunity for all of us.

So when these 4 billion pocket-sized supercomputers visit the vast network of information we've built over the past 30 years, what kind of experience are they getting?

There's a high degree of likelihood that when they visit one of the many websites out there, they'll start by encountering something like a newsletter sign-up dialogue, or maybe a confounding cookie consent policy, perhaps a full-screen app download interstitial.

If they're really lucky, they might get three app download prompts, or two with a cookie consent dialogue and an account sign-up prompt, or maybe one of each with a fixed position ad thrown in for good measure.

Now before you say, that's only content publisher sites that do that, let's look at an example from e-commerce. Now you might argue, well that's just in the US, right? Doesn't seem that way. Or it's not a problem for big companies. Eh, it seems the resources they have enable them to do this type of stuff as well. And while we all probably dislike encountering webpages like this, most of them are designed this way because these techniques work, right?

I mean, app install ads. They happen everywhere. They're all over, from e-commerce to publishing and beyond. While they come in many shapes and sizes, they're there to get people to download native mobile applications. So how effective are they?

Well as usual, it depends on what we mean by effective. It turns out, if you put a full-screen interstitial with a big button in front of people, some portion of them will click. In this example we published from Google, about 9% of people click the big Get the App button. Success, right? It's a pretty good conversion right there.

But the other half of the story is that when we tested removing the full-page interstitial, app downloads only dropped 2%, but daily active users on the web went up 17%. What's the last A-B test you did that had that kind of impact? Perhaps removing things is a great idea for many of our tests instead of just adding stuff.

So if you're measuring active users instead of conversion on app install button clicks, the definition of what's good quickly changes.

When we observe people using our sites, we find app install banners can also have a lot of negative, unintended consequences. In this example, this user is trying to purchase some rosy pink shimmer. And though they've already selected the product they want, they can't seem to find something really important. So they scroll down, they scroll up, they begin to scroll to the left and to the right and back again, searching for that elusive Add to Cart button.

After all, once they have a product they'd like to purchase, the next step is actually checking out. But try as they might, nowhere on the screen is an Add to Cart button to be found. Scrolling doesn't seem to turn it up. So where could it be? Going down again, and down further, coming back up, still nothing. You'd expect it to be somewhere right around here, wouldn't you?

Perhaps they'll tap the little almost cart-like icon at the top. No, nothing there either. Well coming back again, perhaps they'll be able to find it. Let's see how that works. No, still not there. Nothing in the cart, nothing on the page. Out of desperation, what this person decides to do is tap the little X down by the Sephora app ad. And there, lo and behold, an Add to Basket option.

In examples like this and others, app install banners were the direct and sole cause of shopping cart abandonment. In Baymard Institute's testing, 53% of sites displayed one of these banners.

Here's another example. Let's say you want to take a look at this shelved ottoman a little closer. So you tap to zoom, and then you, well, unless you close the app install banner, you can't actually get back to the page where you purchased it.

Which again, if you ask most e-commerce company what metrics they care about, sales conversion is pretty high on the list. So having a user experience that negatively affects that seems like a pretty big deal. And as a result, it's probably worth asking, how do we end up with issues like these?

How can these prevalent app install banners be the direct and sole cause of abandonment when abandonment is the opposite of what we're looking for? Is this a user experience design problem? Maybe it's because these companies aren't investing in user experience.

But when I did a quick job search, I found that not only do they have user experience design teams, but pretty much all of them tout the importance of user-centered design on their business in these listings. Not only that, the job descriptions are filled with superlatives on the impact and importance of great user experience design. So what's going on?

Because like you, I don't find these web experiences to be very user-centric at all. In fact, I'd characterize most of these as user hostile. And it's not just these companies. And I really don't want to single out anyone in particular. But you don't have to look very far to encounter these kinds of experiences on the web today.

Often I hear all this is because business goals are outweighing design goals. PM made me do it. Legal made me do it. But as we saw with app and initial banners, important business goals like daily active users and e-commerce conversions are taking a hit with these approaches.

So why would the business side of the team be forcing us to do this stuff?

To answer this question, we're going to have to go back in time a bit to the Rhode Island School of Design. If you've ever taken a design class like the ones at RISD, you've experienced a design critique. This is where a class goes over people's work and talks about the choices they made and hopefully offers constructive feedback to improve those choices.

One of the things most design critiques have in common is that they're long and they take place in design studios. Most of the students sit on the hard floor or on uncomfortable seats for hours. At RISD, one student noticed this and realized many of his classmates' butts were actually quite sore by the end of these sessions. This is him over there, 26-year-old Joe Gebbia.

Joe experienced this pain firsthand, which led him to search for a solution. He designed a cushion and named it Crit Buns to tackle the problem and plunged full-time into his new business after graduation. Lots of skeptical retailers rejected his cushions until he finally sold 200 units to the Museum of Modern Art in New York, after which orders took off.

Joe was able to build a very user-centered product that addressed real pain points because he was the one experiencing the pain. His butt hurt like his classmates' butts, so he had an intimate knowledge of the problem he set out to solve.

Later, when Joe was in San Francisco, he was encouraged to start another project based on his experience with Crit Buns. He convinced his classmate from RISD, Brian Chesky, to move up to San Francisco to work on what's next.

The rent in San Francisco was high, and the two needed some money. So Joe sent Brian an email outlining a plan to turn their apartment into a bed and breakfast during a design conference, but they only had an air mattress. So they launched a site titled Air Bed and Breakfast, which brought in three guests.

They hosted them, made them breakfast, and earned $80 per head. After that first weekend, they began receiving emails from people around the world asking when the site would be available for other destinations, like London or Japan. From there, Airbnb was born.

Today it serves 2 million guests a night around the world. So Joe and Brian weren't the only ones with the problem they tried to solve. Once again, Joe and Brian had an intimate knowledge of the issue. They needed money. They did the hosting. They made the breakfasts.

When the user is the maker, there's no gap between who is building the product and who is using it. So user-centered design is easy, because you're the user. You feel the same pain, share the same perspective, and can address the problem. As a company grows, though, a gap between the customer and the company starts to open up.

When Brian and Joe recruited their former roommate Nathan to help build the next version of the Air Bed and Breakfast website, he hadn't hosted any guests with them. He didn't make them breakfast. He lacked the intimate knowledge of experiencing the problem. So Joe and Brian probably had to direct him and explain why the service needed to be the way they asked.

But we don't just have development teams in our companies. We might also hire some designers to help develop the UI, the brand, and style our sites.

As the company grows, it's likely we need a product management organization to coordinate the work of all these developers and designers, and make sure someone's thinking about the business impact, the timelines, the market dynamics, and more.

And the functions are likely to continue expanding and growing. Maybe a legal department, perhaps a security team, a growth team, or any number of distinct organizations within the larger one.

As we add these teams and they grow, we're creating distance between decision makers and our customers. The parts of the company closest to our end users aren't the ones making the decisions anymore. They're focused on running the company, keeping the organizations in place, and often rearranging those organizations. Leadership focuses on broad topics like infrastructure and portfolio management.

So this growth begins to create a gap. Let's call it the company and customer gap, or the distance between an organization and its end users. Leadership now has many levels between it and the customer, making awareness and insights more difficult to come by.

If you've ever played telephone, where you send messages down the line, you know things get scrambled pretty quickly, especially when the people playing the game have incentives to change the message as it goes down the line. They may adapt it, consciously or not, to suit the resources they need, or their particular agenda at the time. It happens.

And gradually, these competing agendas and perspectives start creeping into the products we make. That's where the company-customer gap starts showing up in product designs.

Let's look at this simple contact us form. The requirements were just to have customers contact us. So we needed a way to get back to them, find out who they are, and give them a chance to voice their message. Simple form with a name, a way to contact them that's flexible, and a submit button should probably suffice.

But once the sales team hears about this, they want to make sure that these leads have more information so they can route to the appropriate people inside of their team. They probably want an address, a city, and which department or subject is most appropriate for which sales rep.

As engineering discovers they need to build this contact form, they make the point that usernames are actually stored with first name, last name, and streets need separate fields for number, city, and zip code. So the requirements continue to grow.

Marketing finds out that we're talking to customers and of course they have some demographic questions to ask so they can segment our users appropriately and send the right messages to them through all their marketing channels. So gender, date of birth, and a toggle to allow those marketing messages to be sent pop up.

Once legal hears about all the information we're collecting, they definitely are going to require terms of use and a privacy policy acknowledgement. All together, it quickly adds up.

And this isn't a new phenomenon. In fact, it was articulated quite a long time ago in 1967 by Melvin Conway. Conway basically says organizations that produce designs are going to reflect their organizational structure in those designs. In other words, everybody ships their org chart.

So as organizations grow, decision making moves further from end users and the structure of our organization starts to show up in product designs. Our org charts sometimes become so evident in the user-centered experiences we make, you can smell the departments on what we ship. And while that's a problem in of itself, it leads to additional problems as well.

Which brings us to gap number two and a chair. With a chair, what it is and why it's there is a very small leap. You see it and you know what it's for, sitting down. There's very little gap between its form and its function. The design of the product immediately fills that gap. So the purpose of the product aligns tightly with its final form. It's designed to be sat in and the intention is clear.

Now let's look at another product experience, mostly because it's been quite popular, scooter sharing. Let's say you come upon a scooter in your town and want to ride it somewhere. What's that experience like? Well, it starts pretty typically.

Splash screen, a tutorial, sign up form, which then continues to permission dialogue, probably some terms and service agreements, back to some more permissions, followed by filling an account with some funds. Once you do that, we are back to another tutorial probably. And just to wrap it all up, let's throw in another permission dialogue. Phew, you're off.

Now there's reasons for each of these things to exist in this design, but are they all really user centric? Because when you add them up as a full experience, it's not hard to see why someone might call this painful. Well, let's look at some of these reasons and how they're part of growing the second gap.

We'll switch to another scooter sharing service, just to keep things consistent. Here, Hello Bike starts with a splash screen, followed by a sign up form, and then terms and conditions, of which you have no choice but to scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll and scroll. Finally, you get to agree to everything you just looked through.

Probably everybody can agree that 15 screens of legal copy isn't the most user centric experience. Even if you make the case that people benefit from knowing what they're signing up for, this format probably isn't the best way to do that. So when I talk to teams about why they would have this type of experience in their product, I usually hear something like, legal made me do it.

While this may seem like a valid reason, it starts opening up a gap between what something is and why it exists. Instead of the form of the product reflecting why it exists, it now starts to reflect some organizational structure. In this case, requirements, which come from siloed parts of the organization, with different perspectives and views of the world.

So the product experience isn't what directly benefits the customer, but instead what legal or PM or fill in the blanks told us to do. In other words, we start doing things for reasons other than our customers.

Let's look at another one of these reasons and through the lens of another scooter sharing company. Once again, splash screen, few permission dialogues and a tour, which is often justified by saying, everybody's doing it. But what does that mean?

Those of you that have worked at a software design company know it's pretty common to kick things off with what's known as a competitive analysis. That is, you look at what other sites or apps are doing for a specific feature, you print them out, put them on the walls and compare what you see.

In the case of scooter sharing companies, we can look at the onboarding experiences of jump, spin, ofo, bird, lime, and we see across most of them that there's an intro tour explaining the service to people. So the result of this competitive analysis is that intro tours are probably a good idea because everybody else has one, right?

But if you actually take the time to test some of these things, like the music service Vevo did, they looked at how people were using their intro tour through user testing and analytics and they found most people were just skipping through the tutorial without reading any of the copy. So if they're skipping this, what would happen if they just got rid of the tour?

Turns out in a 28-day experiment with over 160,000 participants, the total number of people who got into the app increased, removing the tutorial didn't affect engagement or retention metrics, and more people actually completed signup.

You can see similar principles at work in the evolution of several Google products as well. Google Photos, for instance, used to have an intro tour, an animated tour, and an introduction to its Android app. Following a series of tests, the team ended up with a much reduced experience. Away went the spinning logo, the get started screen, the animated tour, all which were sources of drop-off. All that was left was a redesigned version of the turn on auto backup screen, which was overlaid on top of people's photo galleries.

This step was critical to getting the value out of Google Photos. It's a service that backs up and makes your photos refindable easily. Little in the app works without this step. So the team made it the first and only focus of onboarding.

It's a great illustration of the principle of getting people to product value as fast as possible, but not faster. That is, ask the user for the minimum amount of information you need to get them the most valuable experience. In the case of Google Photos, that's turning on auto backup.

As we saw before, when we start to do things for reasons other than our customers, the gap between what our products are and why they exist expands. This time, our rationale is everybody's doing it. It's a pattern. The competitive analysis showed it's widely used. Yet again, we're doing things for reasons other than our customers.

Coming back to Ofo's scooter sharing service, we can see after sign up there's a promo to try their subscription service. However, looking at the design, there doesn't seem to be any way not to take them up on their offer. Tapping, try it free, goes to two paid plan options. But it turns out if you tap the little arrow in the upper left, you get taken to a map where you can unlock a bike ride without the subscription plan. Not very clear in design.

I have no insider information, but I suspect this was a pretty well performing A-B test. Lots of people hit that try it free button. You've probably heard a lot of people talk about the importance of A-B testing and the impact they can have on conversion. But once again, we need to think about what are we measuring?

The classic A-B testing example is changing the color of a button and seeing results. In this example, 9% more clicks. When test results come back showing one item outperformed the other for a specific metric, it's pretty natural to want to implement that. So we make a product design choice because the data made us do it.

Isn't this how we improve user experiences by testing and seeing how user behavior improves? Yes, but it matters how you define and measure improves. Many companies have results that look like the button color example. In isolation, they show great short-term gains. But when you look at the long-term impact, the numbers tell a different story.

Multiple successful A-B tests you'd think would give you cumulative results much larger than what most companies end up seeing. One of the most common reasons behind this is that we're not using tests with enough contrast. Looking at the impact of a button color change is a pretty low contrast comparison.

A more significant contrast would be to change the action altogether, to do something like promoting a native payment solution by default on specific platforms. The reason the button change is a low contrast change is it doesn't really impact what happens after someone clicks on it. They still go into the same checkout flow, the same forms.

The payment method change is higher contrast because it can completely alter the buying flow. In this case, shifting it from a multi-step form-based process to a single double tap with biometric authentication. So one way of making good use of testing is to try bigger, bolder ideas, ones that have higher risk-reward ratios.

The other way of using testing is basic good hygiene in product launches. Using experiments to check outcomes when making changes, adding new features, and even fixing bugs. This gives you a way to measurably vet any updates and avoid causing problems by monitoring and being able to turn off new changes.

Back to scooter sharing to illustrate yet another way we make decisions for reasons other than our customers. In an effort to scale the impact of design teams, many companies are now investing in design systems or common components to make it easy for teams to apply similar solutions. And nowadays, it's common for me to hear refrains like, well, the design is like that because I was just following the guidelines. But pulling a few off-the-shelf design components from a library is not the same thing as creating a good user experience.

For example, JetRadar, a flight search engine, makes use of material design guidelines in their product. They've used material design input fields in the design of their form and material design floating action buttons for their primary call to action.

But you don't have to be a UX expert to see that the end result is not particularly great. Label and input text is duplicated. What looks like inputs are actually hint text. What looks like hint text is actually labels. Elements are scattered across the page. And the primary action, frankly, just looks like a bug.

JetRadar's most recent design is much more approachable to people, though I could quibble with some of what they do. The point is, simply applying a style guide or design components doesn't ensure your product design works well. In fact, it could have the opposite effect.

Now in fairness, material design actually has updated both of the guidelines I showed earlier to try and cover cases like this. Always be learning.

But the point still stands. There's more to making a holistic user experience than applying guidelines to mockups. And while design systems have great aims, they can quickly become another reason for applying a specific solution for the sake of consistency. And as we just saw, just because something's consistent doesn't necessarily mean it's good.

Too often, the reason for making product decisions is about consistency with guidelines versus customer context and needs. It gets worse when those decisions are made without a deep understanding of the problem space.

But whether it's design guidelines, testing results, competitive analysis, or some other rationale, the more product decisions that are made for reasons other than our end users, the more the gap between what something is and why it exists expands.

To come back to scooter sharing one last time. You can come to a scooter and have a pretty clear sense that you can ride it based on the hardware design. The gap between what it is and why it exists is minimal. With software, we really seem to struggle with closing this gap.

So here's a quick way we might be able to address that. Assume you see a scooter and want to ride. The instructions point you to open your camera and point it at the QR code. From there, it's one tap to a native payment solution with some authentication and you're off riding. Effectively, this makes the process much faster.

It's also worth noting that we're using the web to make this happen. No need for a native mobile app.

Of course, we can quibble on how achievable the design I suggested is, but the point is it really tries to bridge the gap between what something is and why it exists by getting you riding a scooter as soon as possible. This is especially evident when you compare it to the first example we looked at. And in fact, most of the examples we looked at.

So how did they all get that way?

When the distance between the company and the customer increases, people start to do things for reasons other than the end user and that creates a gap between what something is and why it exists. In other words, it's evident in the product design.

But why do we care that there's a gap between what something is and why it exists? This is just design nerdery, right? Well we care because it creates the third gap, which really starts to affect the bottom line and growth of companies. With a big gap between what something is and why it exists, people's path to getting its value increases in length and difficulty.

To illustrate, let's look at an example in e-commerce. I ended up on this one because it made a top 10 e-commerce designs list somewhere. When I followed the link though, I only found two things. An app install banner and a full screen email newsletter promo. Not a great start.

So I did what most people do and dismissed the pop-up, revealing a promotional banner, an icon only navigation system, and a feature carousel. Encouraged by how my dismissal of the free shipping interstitial began to reveal more useful content, I tried removing the two promos at the top and something really interesting happened. I got to a list of categories, which doesn't seem all that compelling until you consider the impact of this UI.

In a few tests, Growth Rock compared standard e-commerce feature carousel based layouts with ones that included a few top level categories, making it clear what kind of products are available on the site. The result was a 5% increase in completed orders. Note the metric we're tracking here. Not clicks on the links, but actual impact on meaningful things, like completed orders.

There is also evidence they ran a similar experiment in another vertical, in this case for an ice cream retailer. Listing their categories up front led to a similar jump in category page views and in this case, a 29% increase in completed orders. Another example comes from Google's mobile optimization efforts, where they saw a similar outcome. Edgars is a large fashion retailer in South Africa.

They removed the animated banners, introduced some high level categories near the top of their screen and saw an increase in revenue per visitor of about 13%. So it seems like getting the categories on the site to be more visible is a good idea, especially if we are tracking impactful metrics like sales.

But there's more we can do here to help people get the value of this product and close that third gap. So next we'll tackle the icon based navigation system. It's worth mentioning that even the icons we take most for granted, like the search icon, are not as universal as we'd like to believe. So let's clarify the search function a little bit.

Instead of using just icons for a critical function like search, we're going to be more explicit in our product UI and close the gap between what something is and why it exists, with a search bar. This also gives us a chance to call out popular items and again reinforce what the site has to offer. I specifically call search out as critical because exposing it by default can also help with conversions.

In this case, boosting the number of searches as the conversion rate for users who search is usually higher than for users who don't interact with it, probably because they have specific intent. So now we have a pulled out search area, category links exposed, and well how else can we make it easier for people to get to the value of this product?

It turns out if we drop the featured image, which probably doesn't drive that much in the way of core metrics, we can show some of the actual products this site sells. Imagine that, showing popular or trending products on an e-commerce site.

But let's not just show two, let's center this module to get more content on the screen and make the images run off the side a bit so people can scroll for more, right where the thumb is for easy one-handed scrolling. This puts the ability to browse top products in a comfortable to reach zone on large screen sizes. Should make all our one-handed millennial users super happy. Because they'll scroll.

Pinterest found that even removing core features like the number of pins and likes in onboarding increased the number of photos they could show people at any given time, which increased the likelihood they'd find content they like and thereby become an active Pinterest user. Same principle applies here.

Overall, I think we've made progress on getting people to experience the value of this site a bit more directly. We could do even better maybe by putting the products up top and categories next. The goal is to get people from the state of, huh, I think I want to get out of here, to I get it, looks like my kind of thing, but you may say, Luke, what about that free shipping promo?

They were making a really big deal out of that, so it must be important, right? Indeed, the top reason for abandoning a shopping cart after browsing is shipping costs, taxes, etc. So free shipping is a winner and people should know about it. I'm not against that.

I just contend that there's probably a better time and place for it. Perhaps on the product page or the actual checkout experience when total cost is on most people's minds. The tricky but very valuable thing that makes mobile design hard is finding the right time and place for things.

It usually isn't right away on the homepage with everything. You can do this all day, but I'll add just one more consideration to this redesign. It's quite possible when someone looks at this design, they could say, but what about the brand? Now, I hope it comes through in the fonts, colors, and especially the products.

What people mean when they say that is something more like this, some aspirational imagery that reflects the personality of the company, serves as a hook for people to dive in, like this edgy Mad Max style look. And I agree, our site design is looking a little too plain.

So we can add in some brand imagery to bring back some soul. But even with this addition, I'd argue we still retain a lot of the functional benefits we've been adding or rather emphasizing by removing other things. Just be mindful that the reasons we're adding the brand imagery are tied to customer needs and not just the agenda of some department, like brand marketing. Else you'll end up back at the product experience that mashes up the agenda of multiple teams, which is increasingly the norm out there.

Now I focused a lot on free people, but they're certainly not alone. Looking at a number of other e-commerce sites, you see they're all doing similar stuff. But the end result is our third gap, the gap between people's first time experience and becoming a happy, satisfied customer. Because I alliterated the first two gaps, I had to do the same here. So we'll call this one the first to fandom gap.

What the first to fandom gap effectively means is that when there's a large gap between what something is and why it exists, it becomes much harder for people to get to its value. It takes longer and more people fall off. And as I mentioned earlier, this really starts to affect the bottom line and your growth as a company.

We've talked about a lot of stuff now. So let's try to summarize things really concisely.

When companies grow, decision making moves further away from users. And people within these organizations start to do things for reasons other than the customer, which means a bunch of things get added to the user experience, which get in the way of people experiencing the real value and purpose of what we make.

Kind of a bummer. So what can we do? How do we improve the situation? How can we bridge these gaps and actually deliver user-centric experiences instead of just saying we're doing so and acting quite differently?

We talked about three gaps. So I'm going to talk about three ways to close them.

Since we saw just how related these gaps are, these techniques actually apply to bridging all of them. The first thing we can do is be mindful that these gaps exist. When the voice of the customer is missing in critical discussions, we need to bring it back. When requirements are conflicting with a holistic product experience, we need to push back on them. When our most important experiences are underperforming, we need to learn why. Awareness is the first step to improving the situation.

Next, metrics. I work on company-wide metrics at Google. Why? Because I believe you are what you measure. Spending the time to get the right metrics affects so many things down the line. Let's say we decide to measure app downloads. Well, we start with a small promo, and then we test out another one. Oh, conversions on it are better. Well, we better keep both of them. Then we add another promo, and installs went up again. So why not drop in more?

Ooh, and things get even better when we make them bigger. Pretty soon, you've become what you measure, a giant app install ad. So please, spend the time working through what metrics to measure and why. Real quick, how to choose metrics.

First, we need to decide what change we actually want to see happen in the world. Next, we gotta figure out how could we possibly measure that change. For each of these possibilities, write down what you think is gonna happen if you start measuring it. What behaviors will you change? What actions will you take?

Next, rank that list by where you see the clearest impact. Then start actually tracking data for your top few, and see if it actually delivers the outcomes you thought it would. When you find a few that actually work for you, make sure to regularly and visibly track those.

Finally, and most importantly, spend time with customers. To illustrate this, I wanna come back to the Airbnb story I started this talk off with. Back in 2009, the Airbnb service wasn't growing. At the behest of Paul Graham, Joe and Brian went out to New York and stayed with a bunch of Airbnb hosts. There they saw firsthand the listings these folks had.

The photos made them look quite poor. They realized this was a problem they could fix, so they rented a camera, took pictures of their hosts' homes, and the next week, the revenue in New York doubled. We used to travel and actually stay with our customers, said Gebbia.

It was the ultimate enlightened empathy. You were so close to the people you were designing for that it informed you in a way that you know an online survey never would. Wise words that actually had real impact.

Based on the success of their New York experience, the Airbnb team created a photography program to scale the process, and from there, they were off to the races. Joe attributes all of this to being closer to his customers, which is really the same experience he had at RISD with design critiques and crit buns.

Getting as close as you can to the problem helps inform how to solve for it. It's not just upstart companies that can make these kind of insights happen. When I worked at eBay back in 2004, we launched a program called Visits that got people within the company into our users' homes.

These programs exist across companies, but people get really caught up in organizational objectives, their own workload, or even documents they're working on, and they don't make time. See the first and second gaps we talked about.

There's many ways you can bring user voice into your organization. We could have a whole talk outlining them. At Google, I organized a weekly meeting titled, What Did We Learn This Week? It had leads from engineering, marketing, PM, UX, and more come together for an hour every day to hear what we learned from quantitative and qualitative research across all the products in our group. It quickly became people's favorite meeting. I mean, look how excited they are in this meeting room, right?

Bottom line is there's no substitute for spending time with customers. Do it regularly, do it often. If the word user research or usability or whatever scares you, don't call it that. Just call it spending time with customers.

It really boils down to staying close to the people using your product and making sure your team directly gets that info as often as they can. Because it's only through this kind of direct empathy, through really seeing the world through our customer's eyes that we can make good on the planet-size opportunity that is mobile.

There's four billion networked pocket-sized supercomputers online right now able to access all the experiences we make for them. Let's make the kind of experiences we want to have and we want our friends and family to have.

The seams we talked about today can open up really quickly. So we need to be vigilant.

When companies grow, decision-making moves further from users. People within these organizations, good people, start to do things for reasons other than the customer. This means a bunch of stuff gets added to product designs, which get in the way of people actually getting to the purpose of what we're making.

Know and apply the simple techniques of knowing what's happening, awareness, taking time to set the right measures, and spending time with your end users.

These are pretty basic principles, but with them we can make a better web for everyone and still have successful businesses, as hopefully I've illustrated.

Please help me make this happen. If not for your customer's sakes, then for your own, because we all want a web that we can enjoy.


Making Product Value Obvious

Mon, 04/29/2024 - 2:00pm

The most valuable contribution product designers can make to software is making the core value of a product clear to the people using it. Yet over and over again... this seemingly simple objective doesn’t get met. Why?

Making the value of a product clear through use (esp. first time use) is how to create customers. You can tell people all you want about your product’s benefits through marketing, onboarding, or a rigorous sales processes but it’s only when they experience the value for themselves that things click. They know why your product matters to them and they tell other people about it. In an ideal state, everyone that should (not all products are for all people) be able to get to this state with your product, does.

But life is not ideal and many factors get in the way of making product value obvious to people. So many, in fact, that I put together a 90min presentation on the most common ones. To summarize, an always increasing set of objectives gets in the way: adherence to a design system or technical architecture, the alignment of different team perspectives, mimicking what competitors are doing, and so on. These dynamics end up becoming the requirements that drive a design process instead of a relentless focus on making product value clear.

Many of these objectives drive us toward applying solutions instead of understanding the problem in depth. For instance, I’m sure many people read what I wrote above and thought: “oh he’s talking about onboarding”. And from there come the inevitable splash screens, tutorials, and tours that have become synonymous with user onboarding. It’s a great example of jumping to answers instead of getting so seeped in the problem that the most obvious solution emerges.

"True simplicity is, well, you just keep on going and going until you get to the point where you go... Yeah, well, of course.” —Jonathan Ive

When you go deep into a problem, solutions like tutorials and tours are not what you end up with. You communicate what a product does for people through the interactions, organization, and visual presentation of the product itself. The two become inseparable and the design feels inevitable unlike intro tours that feel bolted-on or like band-aids.

At this point I imagine folks asking for examples. So here’s two short videos from my afore-mentioned Mind the Gap talk. In the first, I redesign the front page of an e-commerce site to make its value clear up front. The process includes both visual and interaction design changes and pushing back on internal requirements.

In the second, I walk through some examples of onboarding and illustrate how the right design for "first use" will differ significantly between products because of their unique value.

Last but not least, how obvious you need to make things is often... not obvious. Teams building products talk about the value they hope to provide all the time. They've all internalized it (hopefully through regular use of the product they're building) but everyone else has not. What seems obvious to someone on the inside is not obvious to those on the outside. So being clearer than you think is needed is, in fact, needed.

I was recently reminded of this when at a Starbucks. Their food menu was labeled: All-Day Breakfast, Anytime Bakery, and All-Day Lunch. I started thinking they probably started with Breakfast, Bakery, and Lunch as labels but customers kept assuming that breakfast was only available in the mornings. So that label got changed to "All-Day Breakfast". But if breakfast is all-day, when can I get bakery items? And the label changed to "Anytime Bakery". At first blush these labels might seem excessively wordy but they're probably intentionally more obvious. And probably more obvious than originally thought necessary.

For even more on making product value obvious, check out my full Mind the Gap presentation.

AI Pin and Personal Assistant Hardware

Thu, 04/25/2024 - 2:00pm

As the ability for computers to interpret and generate text, images, and video keeps accelerating, it's increasingly clear that multi-modal personal assistants are in our future. But the only way to find out what form they'll take is continued exploration. With that in mind, here's some thoughts on using the AI Pin from Humane.

Humane was founded in 2018 -well before there was widespread awareness about Large Language Model capabilities. Kudos to them for having the conviction that capable enough systems would exist to enable a "smartphone alternative". That said, many of the frustrations (short battery life, overheating, poor usability) I've encountered with the AI Pin likely stem from the fact that it's trying to replace, or at the very least not require, smartphone use.

"It feels like Humane decided early on that the AI Pin couldn’t have a screen no matter what and did a bunch of product and interface gymnastics..." -The Verge

I'm not here to harp on the hardware and software issues the device has as you can find plenty of reviews calling them out. That said, there are enough problems to prevent most people from getting to what could be the hero interactions of the AI Pin.

AI Pin has the components you'd need for multi-modal input and output: a camera, a microphone, a display, a speaker, a network connection, and a processor. The camera and mic are even positioned where we intake the world around us: attached to the clothes on our chest vs. tucked into a pocket or attached to a wrist. Only smartglasses or earbuds seem more optimally located.

So what's wrong? For privacy of the people around you, AI Pin's camera takes seconds to turn on and often does so unreliably as touch gestures to activate it go unrecognized. So getting real-time input from the World around is clunky and the e-ink hand-only display used for output is even clunkier.

The audio side of AI Pin works better. Though slow, its ability to answer nearly any question and utilize context like your current location when responding is exactly the kind ambient computing interaction we've dreamed of since the Knowledge Navigator video from Apple. Minus the screen of course.

When I demoed the ability to effectively talk with a search-enabled Large Language Model to my 15 year-old son, he asked: "Can I get one?" When I pressed on why, he said: "It's like having a little personal assistant with me all the time." Which is the future I outlined up front. So despite all it's flaws, the AI Pin has at least strongly hinted toward where human-computer interaction is headed.

Whether that personal assistant shows up in your palm, on your wrist, in your glasses, or attached to your lapel, we'll find out. As I don't doubt more hardware experiments are coming that let us glimpse the future and ultimately live in it. (insert sound of future here).

©2003 - Present Akamai Design & Development.