Theory Behind Making Something People Want
Jun 24, 2020
Companies like Zoom, Instagram, Slack, and Apple have taken their respective markets by storm. There are of course a variety of factors as to why. But one commonality they share is the extreme focus on making important tasks take close to zero time and effort using their products. Even when it’s at the expense of great pain by the product team such as the attention to every pixel at Slack. Or at great financial cost such as improving battery life by a few minutes through multi-billion dollar R&D budgets at Apple. These companies won people over by choice in heavily saturated spaces. Markets that numerous giants operated in long before them. So how did they know what to build and when to keep investing?
While analyzing 10s of products and iterating with 100s of millions of users at companies like Microsoft, Amazon, and smaller startups I’ve developed a battle tested theory of how. A lot of ideas are cross applied to product from statistics, machine learning, and AI since that’s what my background is in. By putting this theory to practice you can assess exactly how every change within your product affects the various markets your customers are representative of. It also lets you evaluate if your product is easy to use with quantitative certainty.
It’s important to point out the nuances of this approach and what it doesn’t do. Not all products have to be easy to use to build a company around them. Look at Salesforce and any IBM product. They just have to make important tasks easier than today for people you know how to sell to. Also, there’s still no shortcut to figuring out what to build other than qualitative research with real people, probing markets with an MVP, and learning about a domain.
What this theory does help you with is evaluating the worth of your solution within different markets that use your product. It can also inspire ideas by highlighting what your existing user base already “hacks” with your product. Lastly, by observing exactly how various personas in your product react to changes you build you can understand what’s important to them so you know what areas to focus on.
I, along with my 2 co-founders believed in this theory so much that we all quit our $200k-300k jobs at Microsoft over 2 years ago in order to productize it at ProductMarketFit.AI. Our goal is to make it accessible to everybody including side projects, startups, and consultants without requiring hiring a 25 person data science team like larger enterprises do.
I’m going to split this post into 2 parts. Part 1 is focused on:
- Building survival vs. pleasure technology
- Evaluating ease of use quantitatively
Part 2 is focused on:
- Evaluating the importance of tasks for your user base
- Probing markets with your product
I’m always growing and learning so I look forward to all types of discussion. Feel free to reach out to me on any platform. Now to it.
Survival & Pleasure
The reason humans do anything roughly falls into two buckets:
We brush our teeth, earn money, and sleep so we can continue pain-free existence. When there’s time to spare we long for love, revel in our accomplishments, and get dopamine hits from entertainment.
When negative experiences such as a crappy boss, loneliness, boring tasks, and an aching back invade our experience we lose pleasure in life. Small pains grow into big ones through repetition and large pains can quickly eat away at our happiness.
Happiness over time
In one statement:
Humans strive to maximize the longevity of their positive subjective experience.
Building Survival & Pleasure Technology
The point of technology is to make survival easy and maximize pleasure. You can roughly map market categories this way:
- Survival: Prosumer, SMB, Enterprise
- Pleasure: Consumer
Note: There’s no mainstream professional market in pleasure yet. Are there not people who love their work? Survival is hardly the word that comes to mind of why Warren Buffet works. People who invent interesting technologies or professionally help others in non-profits tend to highly enjoy their work. My theory is it’s too small of a market today. An alternative perspective is the enjoyment that comes from such work is not during the work but after.
If we look at the macro goal of products in the survival category they:
- Collapse extremely repetitive or unpleasant tasks important to survival.
- Expand what is possible for one person to do.
Example: Each one of us used to plow fields for 16 hours a day to eat. Now, one person with a tractor does that for tens of thousands of us.
Because of the innovations of the past most of our jobs are much easier today. Check out Chapter 5 in the book Sapiens to read more about the history of this. We also have more time leftover for pleasurable activities. Let’s break down the collapse and expand loop of survival technology in detail.
Collapse & Expand Loop
Take a bunch of survival tasks:
Some of these tasks are pretty painful so we dedicate people to them. Other tasks are still important but are easy enough for the same person to handle.
We begin the work to break these tasks down. They are not enjoyable so our ultimate goal is to eliminate. En route to that goal we:
Technological paradigm shifts enable technology to be created. E.g.: The discovery of electricity, transistors, internet, and cell phones. Each paradigm shift makes it easier & cheaper to enrich more tasks with technology. If the real world task has enough scale and importance to sustain the technology, companies form around it.
Once enriched, through intense competition and research, companies simplify the technology to be easier to use. With time, tasks that used to be hard become easy. Some tasks are automated away.
Once a set of tasks become automated or easy we eliminate the need to occupy multiple people’s time and effort with them. E.g.: What used to take 3 people can now be done by 1 person.
Through more intense competition an all-in-one solution usually emerges that collapses tasks within a similar activity into one product. This lowers the mental overhead of this activity overall for the person. The new concepts and work do mean the overall task becomes pretty difficult for one person initially but with time it will enter the collapse phase itself again.
With 2 people freed up, they can go on to tackle other important difficult tasks of the world.
There are a lot of questions and concerns around:
- Do individuals get to benefit from inventing collapse technology forever if they secure it through a monopoly?
- What should happen when there is enough technology to ensure survival for everyone but it’s owned by a tiny % of people?
- What if the tasks left are extremely complicated and out of reach for most people to learn how to do?
I’m aware these concerns exist but it’s out of scope for this post. If you are interested in the topic, Kurzgesagt did this excellent video about it.
While the demands of survival are clear, how we derive pleasure is much more varied. It is the long tail of wants. There are of course obvious ones which directly tie into our visceral pleasure center such as food and sex. I would argue they make up a relatively small portion of our time. A much more time consuming part of our existence in this category is dealing with the emotional soup that is our subjective experience.
Our subjective experience is influenced by genetics, environment, actions, and beliefs. Internal ego drivers vary drastically and whether a solution is able to deliver on ones you enjoy is hard to predict. It’s also very hard to know how many people share that unique set of circumstances and desires to determine whether it can sustain a successful company. That’s why you see a bunch of weird consumer startups pop up all the time.
The successful consumer products usually identify an often contrarian chunk of this emotional soup through a stroke of genius, luck, or personal experience. Products that find an initial way in then iterate to retrofit their product more and more to the internals of how our minds work. If they are able to do so repetitively, they stay in business and grow.
It would be silly to try to summarize a framework for building pleasure technology. How big is the market for each one of these and how would you create something that makes you feel these?
- Aesthetic Appreciation
And in one statement:
Technology collapses the time & effort required to survive and opens up new ways to derive pleasure.
The Path Of Least Resistance
Survival and pleasure technology do have one thing in common. You want to eliminate as much as possible of what stands in the way of the outcome the person is there for. Let’s zoom into that aspect of the user experience (UX) and user interface (UI).
Visual Design (VD)
We can intuitively tell what looks good but few people know how to create it. Visual design forms the first impression of your product. Studies have shown that we have 50 milliseconds before users have made their first judgments of the product. If you can’t convince someone you’re credible you might never get a chance to solve their problem.
Here is a list of a bunch of visual design concepts. Key ones are:
How to do it well
There are some popular methods that allow the evaluation of visual design:
- Have people pick between 2 designs.
- Ask people to rank designs from most to least favorite.
- Aggregate asking on a scale of 1-7, how beautiful do you find this design?
Good old fashioned guerilla research studies on the street work quite well for this one. There are some cool sites like UserTesting and a slew of others if you want to “get out of the building” remotely.
You can also choose to rely on the intuition of a highly trained expert, aka a designer. Ideally, you want to get the response of people in your target market to make sure your VD triggers the right emotional response for them.
Interaction Design (ID)
ID’s what’s behind all those clicks, drags, and swipes. Humans have a limited amount of mental energy per day. Pleasurable tasks replenish it while other tasks deplete it.
Taken to infinity, you want non-enjoyable tasks to be eliminated. When a task is fun you want to minimize the interaction needed to get to the fun part.
How to do it well
Obviously it’s not always going to be practical to eliminate all interactions. It’s also hard to know what is minimal “enough” to make people happy since it depends on their perception of effort and time.
What you can do is measure how long each task in your product takes and how much interaction it requires. When you make a change, you want to ensure that for each task and subtask the amount of interaction does not go up.
In practice these metrics looks like this:
Task 1 Averages
Time to success (tts): 54 seconds
Interactions (click, drag, scroll): 17
Keyboard Inputs: 24
Task 2 Averages
Time to success (TTS): 4 minutes
Interactions (click, drag, scroll): 192
Keyboard Inputs: 2
Every time you make a change you can tell if TTS & interactions are staying the same or better yet going down.
It can be hard to define what a task is in traditional analytic tools. Google Analytics and Amplitude give power user interfaces to define “funnels” of usage but it’s very hard to capture the variance of how people accomplish tasks, the loops involved, and so on.
That’s one of the pain points we built PMFIT.AI to address. We use AI to assist capturing the 1000 of sub-tasks and larger tasks within your product so it becomes tractable to measure and easy to understand for the entire team.
Information Architecture (IA)
Various features are used to accomplish specific tasks. How they are laid out on pages, panes, and modals should reflect those tasks. Good IA lays out components that are relevant to the task at hand. All extra information is waste.
Our attention is a limited resource. The more information we have to mentally filter to achieve a task, the less mental energy we have left afterwards. User experiences where just the right components are there at just the right moment are the best. They feel light and we have more mental energy in order to get more done.
How to do it well
Every time someone performs a task in your product, count the number of pixels filled with information that they didn’t use. If you identify information that is rarely used, consider:
- giving it less prominence,
- moving it to a different space or,
- removing it altogether.
If certain features are always used in a certain order, move them closer in proximity. If a certain task requires traveling between 4 screens, consider if it would be easier mentally to collapse it into one screen or split it between more screens.
There’s the conflicting problem of discoverability. How can you know when someone would highly appreciate a feature to accomplish their task but only once in a while? One example most are familiar with is flight check-in notifications. It’s something I only really need to remember about a few days before my flight and then a few hours beforehand.
In practice this looks like this:
Task 1 Averages
Unique surfaces seen: 2
Times surface seen: 4
Unique elements seen: 12
Unique elements used: 4
On-page distance between elements used: 1100px
Scroll travel: 9000px
Mouse travel: 17000px
Feature 1 Averages
% of the time used when seen: 1%
% of users ever used: 9%
You can use this information to figure out how you’re doing with your tasks:
And then A/B test to ensure those tasks are being completed just as often but with less steps and less information overload:
You can get a sense of how this looks on your site using tools like Hotjar and Fullstory. Because this takes an incredible amount of time and energy, especially if you have more than a 1000 users, we built ProductMarketFit.AI to auto-metrize this as well.
You can see where there’s a lot of unnecessary mental overhead for a lot of users and THEN watch videos to get a few examples of the issue at hand. When you A/B test changes you can assess how these metrics changed and then inspect the metrics and corresponding videos of A versus B.
Conceptual Design (CD)
Last but not least is CD. Good CD matches real world tasks you are familiar with to the tasks your product improves. If done well, the entire product can feel intuitive without ever having seen the product.
When the task at hand is finding directions, it would be pretty strange if it was a list. And if you’re getting groceries, it would be strange if the checklist was on a map. Conceptual design is about ensuring the tasks your product helps with fit into the user’s brain naturally.
How to do it well
The easiest ways are:
- Ask people how they do your task today.
- Pit concepts against each other in a user study to complete the task.
- See how tasks in similar domains are solved today.
You can use the same tools that were listed in the section about VD as well to speed this up if you can’t get a hold of people in person.
Building the path of least resistance
To recap everything, I’ve made a mini graphic of each piece of the UX/UI pyramid. It was once only possible to measure this sort of thing in a lab once in a while. With new user research tools you can quickly evaluate VD and CD remotely, and with tools like PMFIT.AI you can now baseline ID, IA and measure the delta with every change for your entire audience or different segments.
The Entire Solution Space
Making the UX and UI amazing only makes sense for things that are important. Which brings us to our next step, the feature set:
When deciding on which feature set to implement, figuring out how difficult and important the task is today in the real world is crucial. I’ll continue with the new ways to measure and track the feature set in Part 2. If you are interested in being notified of new posts, please sign up by email here or follow me on Twitter/LinkedIn.