Tag: Data Science

The slipperiness of UX data

Posted on

In my article proving design, I talked about how hard it is to have proofs for making the right product or product decisions. Some projects are so expensive that it takes a lot of convincing to get a budget for them. It’s a natural trade-off. It doesn’t get easier after you’ve done parts of the project, or even after you’ve done the project and are now interested in moving forward with a second stage of development.

Many UX professionals talk about the importance of data but let’s be honest, in the cycle of design and decision making there are countless things that cannot be measured.

What can be measured?

  1. Do people need your product?
  2. The product itself and how people use it.
  3. Ideas, and iterations — Using user research.

Basically everything you can do with your team. It adds up to around 20% of the creation process.

What do you create? What informs your ideas? Are you influenced by other designs you can’t measure? Hell yeah!

What cannot be measured?

Competitors

You can’t know why your competitors behaved the way they did. You don’t have their data and you can’t know their decision-making process.

Pre-product behavior

There are many marketing products that are trying to solve this. However, in this part of the user journey, the designers have zero control. The user journey is driven by the facilitators whether it’s the OS or the platform. Each platform will supply you with some data and measurements but it’s not exactly monitoring UX, it’s more generic and marketing led. In big organizations, it’ll also be a challenge to get these data points. In addition, every piece of data should be verified. With platforms, it’s almost impossible to verify.

Analytics

In your service, you’ll need to check and correlate through different tools (MixPanel, GTM, Data studio, etc.). Understanding analytics tools have become essential for UX and Product roles. This is how companies make crucial decisions and that’s why it is checked and cross matched, usually with three to four systems, to compare and see if the data is reliable.

OS design patterns

The fact that Google decided something should look the way it does doesn’t mean it’s the best way. It means that they probably measured it and it works. It also means that with their level of influence, many apps will adopt it and it’ll become familiar. But it doesn’t necessarily mean it’s better. Some of these decisions are made to differentiate from other platforms like iOS or Windows. Other decisions are a compromised solution to a great design because the design might be patented. That’s precisely why Google and Microsoft bought Motorola and Nokia, stripped them of their patents, and then sold them on to someone else. So if you’ve seen a design, even if it’s famous, it doesn’t mean it’s the best practice.

Just because it works it doesn’t mean it’s a nice experience. Many companies don’t see a reason to change. It’s very common when a company has a monopoly.. For example buying stuff on eBay…does it work?, Yeah…Is it a nice experience? No. Everything is cumbersome: receiving messages, sending, going through versions of eBay from 2000 till today.

It works but it ain’t nice and at times very confusing

In comparison, Amazon are more ambitious and “very slowly” redesign their experiences to be functional and delightful.

You’ve got the Data! But, wait, it might be skewed.

Let’s have a look at how this can happen.

Wrong implementation

Just a simple line of code or a selection of a wrong event could cause every click to be counted as two. That’s why it’s important to check with multiple systems — which, as I mentioned earlier, could be problematic at the stage prior to the user journey beginning in earnest.

Intentions

Even if you have a lot of verified data, how can you believe the data that you see? Every person that collects data (arguably even scientists) is trying to show the data in a way that will flatter their agenda. Data can be collected and presented in a non-neutral way. It’s natural and happens with everyone from marketing companies to UX designers who just want their projects to be successful.

Source giphy.com

Presentation

The medium is the weapon and it’s important to understand why something was chosen, in a similar way to understanding graphic design decisions: What do they show me, and what don’t they show me.

A few current examples of skewed data and how it has been used:

  1. Facebook admitted to having wrong measurements for the 10th time
  2. Facebook is accused of being part of the Fake news problem…Google is too, but it’s used much less for leisure and content consumption.
  3. Facebook is deleting tens of thousands of Fake users, which is why they keep tweaking the news feed, and Google is doing the same with search results.
  4. Cambridge Analytica is suspected of and grilled about their methods for influencing users in the UK / US.

Key ways to deal with it

Critic

Be harsh and critical, try to look for the angle. Life sucks if you always think everyone has an interest but even though awareness drives sadness it’s smarter to look at things critically, especially in business. So when you see a new feature, after you finish getting excited or booing it, think about why they created it? Whose decision was it to make it and what’s their interest? Link its value to the business, marketing, user satisfaction, design etc. Guess which department came up with this concept. Think about where they could take it to next. What’s the future of it?

Influenced but aware

There is nothing public, what you have is a trail of user experience data. I got responses for a previous post I wrote about Facebook that said: “But where is the data?” The answer is: This data is internal and not available to anyone else. It’s too secretive to expose, it’s their secret sauce. Does that mean I’m not allowed to write about or analyze it? I don’t think so.

In Instagram, you’d know how many pictures are uploaded to Instagram because it’s a financial data that affects retention/time spent. But you wouldn’t know how many of these pictures are uploaded from a computer, user’s gallery or a professional camera. It’s just important to accept it when critiquing or being influenced by it and to know the limitations of the data you’re dealing with.

Here is an example where you have data, but can only see part of the picturee: “Apple’s revenue from repair is bundled in with its “services” revenue, alongside digital content sales, AppleCare, and Apple Pay revenue. While there is no good way to figure out how much revenue comes from repair, Apple’s services revenue pulled down $7.04 billion in net sales, out of $52.90 billion total.”

Hack

Be aware of your level of control, but see if you can take it further. The difference between owning an OS and participating in one is huge. When I was working for Samsung we were designing the core of Tizen OS for TV. We had control over everything without limitations. We could track everything we wanted to if we built it. But when you are a part of an ecosystem you need to play by the rules and get whatever you can throughout the process. That’s why designing for a native OS is such fun, especially if others are building and increasing your knowledge.

Data is important, but I would argue that decision making can only be done based on it to a certain degree. In my opinion, around 70% of what constructs the decision is experience, aspirations, and alignment with the other sides of the business. A good designer or product guy should influence and convince but it’s not all up to data. Data is just another tool in the arsenal and it’s good for specific use.

Reimagining storage

Posted on

Storage is everywhere

Look at your house, half of the things there are for storage. Sometimes there is storage for storage like a drawer for small pots that are placed inside pots. In the computer science world they took this metaphor and used it in order to help people understand where things are. Throughout time, new metaphors came like search and ephemeral links. When I was 21 I taught a course called “Intro for computers” to the elderly community, in it I remember explaining how files are structured in folders. I could see how the physical metaphor helped people understand how digital things work.

Nest Storage by Joseph & Joseph (Bermondsey, London)

Unlike the organisation of physical objects in a house, it’s harder to anticipate where things will be in a computer if you weren’t the one who placed them there. Then search came, but still nowadays search is somehow so much better for things online in comparison to our own computers. A lot of research has been done about content indexing for the web. Obviously it was driven by search engine optimisation and the business benefits you get if you reach the top three in the search results page. However, people are not familiar with these methods and don’t apply them to the way they work with files on their devices.

In iOS, Apple chose to eliminate the finder and files in general. Apps replaced icons on a desktop, the App Store replaced thousands of websites and the only folders we actively access on the phone are gallery, videos, and music. Even Android concealed the folders and through Share menus users can probably avoid the transitions that were a norm on computers.

How do we deal with files nowadays?

  • We sometimes look them up via search
  • We put them in folders, in folders, in folders
  • Files sometimes come as attachments to emails or messaging apps.
  • We (if advanced users) tag them using words or colors

So here are the problems:

  1. Search — It exists but indexing needs to happen automatically.
  2. Self organisation — lacking and not automated in any way.

A new horizon for our contents

Companies see value in holding on to user data, it’s of course stored on the cloud. There, companies can compare, analyze it, do machine learning, and segment it for advertising. Their ability to centralize things helps them to give us a smoother experience by creating an ecosystem.

But storage isn’t just a file or a folder. It has its physical presence, as a file but it could also be an object in real life, a part of a story, a timeline with interactions, challenges and emotions. How can we rate the importance of things and interlink them into past events? How can people relive their day a while ago, black mirror style.

The average person has x amount of files. That is probably x photos, x videos, etc. For many of these there are dedicated software. For music you have players, for photos you have galleries with new search features and event creating features, for video you have your library or Netflix. But what about the rest of the files? What about you Excel sheets, your writings, your projects?

Think of the revolution that happened in the past couple of years around photos. We started having albums, then stories and events. In addition to that we have the elements of sharing, identifying who is there, conquering the world with our pins. In addition to that we have another layer of filters, animations, texts, drawing, emoji. All that is even without speaking about video which needs to adapt these kind of paradigms to create meaningful movies…it’s nicer to share.

If we are looking at the work area there are products that try to eliminate the file by opening different streams of information like Slack or hipChat.

The one thing that these products still lack is the ability to convert them all to a meaningful project, which is attempted by project management software. Project management software tries to display things into different ways which helps covering use cases and needs. However, still most of its innovation is around aggregation, displaying calendars, tasks and reports. Things get complicated when tasks start having files attached to them.

What kind of meaningful file systems can we create?

 

Imagine the revolution that happened for photos happening for files. Photos already have so much information linked to them like Geolocation, faces, events, text, and the content in it. We finally have pieces of software that has the ability to analyse our files. A good example is Microsoft LipNet project that extracts subtitles from moving lips in video. Imagine the power of AI working on your files. All these revisions, all these old stuff, memories, movies, books, magazines, pdfs, photoshop files, sheets and notes.

How many times have you tried to be organised with revisions, your savings of docs? How many times you’ve created a file management system that you thought would work? Or tried a new piece of software to run multiple things but soon after realised it doesn’t work for everything? It is a challenge that hasn’t been solved in the work environment or the consumer environment.

Some products that are focused on the cloud started implementing seeds of this concept. It usually comes in the form of version control and shared files. It is amazing that people can now work collaboratively in real time. However I think it will spiral to the same problem. There are no good ways to manage files and bring them together into meaningful contextual structure that could once and for all help us get rid of nesting and divided content.

In terms of users navigation there are currently five ways to gain content on the web:

  1. Search — have a specific thing you’re look for
  2. Discovery — get a general search or navigation and explore
  3. Feed — Follow people
  4. Message — get pushed from outsiders
  5. URL dwelling and navigation — repeated go in and out

But on the computer with your own creation there are only two

  1. Search — for specific file and then maybe it is associated by folder
  2. Folder dwelling — No discovery, just looking around clumsily for your stuff and trying to create a connection or trying to remember.

It’s an early age for this type of service offers

There are some initial sparks of innovation around systems that simplify files. For example looking at version control in Google Drive or Dropbox. The idea that a user can work on the same file with the same filename without even thinking they will lose important information gives comfort. No more file revisions can be a good step in helping us find things in a better way.

Code bases like Github also help control projects with versions control that allows collaborative work, and merging of branches. That is also a good step since it proves that companies are thinking on file projects as a whole. However it still can’t help creating contextual connections or meaningful testing environments.

Final note

AI systems are finally here and the same thing they do on the cloud they can do in our files. We can maybe finally get rid of files the way we think of them and finally be organised, contextual and god forbid, the structure doesn’t have to be permanent, it can be dynamic according to the user needs and the content that is being accumulated every day.