A Perfect Use for Personas

I was reading Dave McClure's post about changes to menus (and its not always flattering Hacker News thread), and I found myself both violently agreeing and disagreeing with both. I kept thinking something along the lines of, "That would be great! Except when it would be incredibly annoying!"

That's when I realized what was missing for me: personas.

 First off, apologies to Dave, who certainly doesn't need me to defend or improve his ideas. This is just meant to be an explanation of the process I went through as a designer and researcher to understand my weird, ambivalent reaction to his product suggestions. Here are the problems that Dave listed in his post that he was solving for:

• Too many items, not enough pictures, simpler & more obvious recommendations. 
• Not online, no order history, no reviews, no friends, no loyalty program, no a/b testing. 
• Have to wait forever for waiter to order, re-order & pay. 
• Nothing to do while I'm waiting. 

Then he presented reasonable solutions to these problems. All of the suggestions seemed geared toward making restaurants quicker, more efficient, and lower touch. Interestingly, both the Hacker News complaints and my own seemed to be from the point of view of people who do not have these problems. They were saying things like, "this would make restaurants awful!" but what they really meant was, "I, as a potential user, don't identify with that particular problem you're trying to solve, so your solution does not really apply to me."

In other words, Dave's suggested solutions might be great for people who have these problems but might not appeal at all to people who don't have these problems.

 So, then I started to think about the types of people who would have those types of problems. I put together a few rudimentary personas of people who likely would benefit from things like recommendations, entertainment while waiting, a more efficient order process, and a faster way to pay.

As a note, these personas are behavioral, not demographic. This means that you might sometimes fit into one of them and at other times you wouldn't. It depends more on what you do than who you are.

The Business Person

Imagine that you're on a business trip to someplace you've never been. You're quite busy, and it's likely that you'll have to eat a few meals on your own, possibly on the way to or from a meeting or the airport. You're not a fan of fast food, so you'd rather be able to find something you like at an interesting local place than at a big national chain.

 In this instance you might LOVE having things like recommendations from people you trusted, pictures on the menu of unfamiliar dishes, and a quick, efficient ordering and payment system that guaranteed you wouldn't hang around for twenty minutes waiting for a bill. You might also really enjoy some entertainment so that you'd have something to do that wasn't stare creepily at the other patrons.

The Barfly

Now imagine that you're at an incredibly crowded night spot. You are desperate for a bourbon, but you don't want to queue up five deep at the bar to try to get someone's attention. You manage to get a table, but now you have to decide whether to leave it to flag down one of the few waitresses or or just wait it out.

 In this instance you would almost certainly be excited to be able to order and pay directly from your table using some sort of tablet. You'd also be able to quickly order your second, third, and (dare I say it) fourth rounds without having to go through the whole process again or count on the waitstaff knowing exactly when to ask if you want a refill.

The Group Luncher

Last one for now, I promise. You're out to lunch with eight of your coworkers. You need to get back to the office in 45 minutes for another stupid meeting. You don't want to spend 10 of those minutes just for a waiter to make it to your table and take your orders. Also, you really don't want to be the one in charge of figuring out how to split the bill, especially since three of your coworkers always get booze, one of them never eats more than a salad, and two of them order the most expensive thing on the menu.

In this instance, you'd be thrilled to be able to just sit down, punch in your order (and your credit card!), get your food delivered to you quickly, and get to spend more time chatting with that cute new person in accounting rather than negotiating who forgot to figure in tax to the amount they owe on the bill. 

And the rest...

There are probably a half dozen other hypothetical persona groups, all of which would obviously need to be validated (or invalidated) with various forms of user research and quantitative testing.

 The persona groups that aren't on this list are also important. Many of these types of innovations might make things worse for the types of folks who are enjoying the experience of being in a restaurant as an event. For example, a romantic dinner for two at a high end restaurant is not improved by shaving thirty minutes off the wait between courses. Other people might enjoy the personal exchange with the waiter or a consultation from a sommelier more than reading about items on a tablet.

That's ok. These products aren't necessarily going to be for every type of restaurant all at once. There's no need to worry that suddenly Manresa is going to be putting pictures on the menu like Denny's.

 The reason I bring this up is that it often helps me to evaluate product ideas through the eyes of the people I expect to use the product. When I find myself saying things like, "Driving sucks! I'm going to fix driving!" I have to step back and realize that driving (like eating in restaurants) is an almost universal activity that has a constellation of problems, many of which are not shared by all types of drivers (or eaters). If you think your startup has a brand new product that's going to solve all the driving problems for stock car drivers, commuters, and truck drivers, I think you're probably wrong.

Instead of arguing back and forth whether or not these problems exist, it's very easy to identify particular types of people for whom these problems MIGHT exist and then do some simple qualitative research to see if you're right. After all, we know at least one person (Dave) has these problems that he wants solved. Presumably Dave (or the companies he invests in) are doing the sort of research necessary to make sure that there are enough people like Dave to make a profitable market. That market might not include you, but there are lots of wildly successful products you don't like.

 So. Long story short: personas, yay!

Postscript

For those of you who notice these things, you're right, I didn't include the personas for the other side of the equation: the restaurant owners. Whenever your customers (the people who give you money) and your users (the people who actually use your product) are different, you're in a much more complicated space from a user experience point of view. I'm assuming that, if we can make a specific type of end user happy enough it will make the types of restaurant owners who cater to those users interested in purchasing the product.

 That's just another hypothesis, and all hypotheses need to be validated, not assumed to be facts.

Startups Shouldn't Hire User Researchers

Everybody seems to think I'm a user researcher, which is not strictly true.

It's true, I write a lot about user research, and I've certainly done my share of it over the course of my career. But I don't really consider myself a user researcher. I do enough user research to be extremely effective at being an interaction designer and product manager.

There are lots of actual user researchers with degrees in psychology and specialized training who are doing much more interesting and complicated research than I do. I respect those people. They are scientists in many cases. But I don't think that most of them should work for startups. In fact, I don't think that small startups should ever hire a user researcher just to do research.

Don't get me wrong. User research is critical to startups. How else are they supposed to understand their potential customers and find product market fit?

No, the reason people shouldn't hire people to do their user research is that learning about your customer is the single most important part of your startup. If you're outsourcing that to a person who isn't directly responsible for making critical product decisions, then you are making a horrible mistake.

I see startups do this over and over. They hire a consultant, or even a regular employee, to come in and get to know their users for them. That person goes out and runs a lot of tests and then prepares a report and hands it over to the people in charge of making product decisions. Half the time the product owners ignore the research or fail to understand the real implications of it.

And why wouldn't they? The product managers weren't there for the process of talking to users, so they almost certainly haven't bought into the results. It's really easy to ignore a bunch of bad things somebody else wrote about your idea in a Powerpoint presentation. It's a lot harder to ignore half a dozen real users saying those things to your face and showing you problems that they're having in real life.

The right way to do research in a startup is to have the people who are responsible for making decisions about your product intimately involved in the research itself. That means that product owners and UX people are designing and running the tests. Even the engineers should watch some of the sessions and hear first hand what their users are going through.

The reason I talk so much about user research is that I want you, the entrepreneurs, to learn enough about it so that you can DO IT YOURSELVES. You're welcome to hire people like me, or even real user researchers, to teach you what you need to do. But having somebody else do the research for you is not an option. At least, it's not one that you should use if you're still trying to find product market fit or learn anything actionable about your customers.

Stop thinking of user researchers as people you hire to get to know your users, and start thinking of yourself as a user researcher. At startups, you should all be user researchers, especially if what you really are is a designer or product manager.

Here's the Problem With Your Product

As I mentioned on Twitter, I often answer emails that people send me asking questions about UX. I enjoy it. It helps keep me in touch with what type of questions entrepreneurs are having about their products.

Whenever I mention that I'm happy to answer UX questions (for free, guys! Seriously. I have a book to procrastinate, after all.) I tend to get one particular question over and over. It is some variant on "How is the UX for my product/site?"

I'm publishing an answer that is very similar to one I recently sent to a nice entrepreneur who asked me this question. I'm doing this because I basically end up writing the exact same thing over and over when people ask me this question, and I'd love to get some different questions. Please note, unless you are building one of a fairly small number of products that I use on a regular basis, this answer applies to you.


I can't give you insight into your site, because I'm not the target customer. If you ask for my opinion, it's going to be mostly useless, because it really doesn't matter what I think about your product. It matters what your user thinks about your product.

It's like if somebody asked you about your opinion of their spaceship. Presumably you don't fly spaceships, so your opinion is almost certainly not going to be super relevant to interspace travel methods. You want feedback about spaceships, you ask an astronaut or an extra terrestrial (no, I do not have suggestions on recruiting for that study).

In order to get in touch with some of your users, I'd recommend that you do the following:

Figure out exactly what you are concerned about with your site or product. 

• Do you want to know if new users understand the messaging?
• Do you want to know how people are finding specific information or performing tasks?
• Do you want to know the general behavior of people coming to your site?
• Do you care about the experience of current users, new users, returning users, etc.?
• Do you care about what the look and feel of your site is telling new people? 
• Are you wondering why your revenue is too low?
• Are you concerned that people aren't coming back? 
• Do you want to encourage people to share more? 
• Are you having trouble converting free users into paying users? 

Figure out which metrics you care about that you'd like to change, and do some validation around why they are what they are. 


For example, if your conversion is too low, you're going to need to figure out if people don't want what you're selling, don't understand what you're selling, or don't care enough to pay you for what you're selling.

Based on what you want to learn, you need to find some way of learning that. You can ask me for specific advice on those sorts of things. The more specific you are about the type of user and the type of thing you want to learn, the easier it is for me to suggest doing something.

You can also ask me for advice on things like what to do when you've found out that people aren't sharing because they don't understand how to do that. Or if you've learned that people aren't converting from free memberships because they're not understanding the value that they'd get from your paid product. In fact, I'm happy to give you advice about how to proceed with your UX for anything that is at this level of specificity.

There is no such thing as generic "UX". Your user experience only makes sense in the context of your particular users, what their behavior is, and what you want their behavior to be.

When Talking to Users Saves You Time

I mentor a few young designers, which is great, because not only do I know exactly who I want to hire when I’m building a team, but they also share interesting stories about their current companies.

I was speaking with one of them a couple of weeks ago, and she shared a story that sounded incredibly familiar. I think this happens to all designers who work with a sales force at some point.

 The designer, whom we will call Jane, is working on the user experience for an enterprise product for hiring managers. The product has some competitors that have been around for awhile.

One day, a few weeks back, the sales team came to the product team and said, “We need Feature X. All of our competitors have Feature X, and we’ve heard from some of our potential customers that they won’t buy us if we don’t have Feature X.”

 Jane and her team looked at the competition’s implementation of the feature, which had a lot of bells and whistles. The product team asked sales which parts of Feature X was most important to the potential customers. “All of it,” sales replied.

Jane’s team starting pushing back. This was not a simple feature. They estimated that it would take months to get the feature to be comparable with the competition. There was one part of Feature X in particular, the live video part, that Jane knew would be incredibly tough to design and build, simply because of all the implicit requirements that would make it useful. They explained this to the sales department, but the sales department continued to complain that they couldn’t do their jobs without Feature X.

Finally, Jane insisted on speaking directly with a customer. A meeting was lined up with a few representatives of the company. Jane started off by asking how the potential customer would use Feature X. They gave detailed explanations of exactly the places that they needed Feature X, none of which had been conveyed by the sales team.

Interestingly, none of the uses they had for Feature X involved the live video part of the feature that was worrying the product team. Finally, Jane came right out and said, “Tell us about live video. How do you feel about it.” The potential customers shrugged. “I guess it might be useful,” they said. Jane asked, “Would not having live video prevent you from buying our product if we had Feature X?” “Not at all,” the potential customers said.

Jane’s team then had a similar conversation with other customers and potential customers. The product team gladly put the much smaller Feature X, minus the expensive live video feature, onto their product roadmap. They also left out a few other parts of Feature X that didn't solve actual user problems, and created a design for Feature X that was significantly different from the competitors' versions but that addressed all the customers' issues.

Sales was happy because now they could tell potential customers that they were competitive on features. Jane was happy because she was able to quickly identify a real customer problem and solve it, rather than fighting with sales about something that would take too long to build and would include features the customers didn’t actually want.

 My prediction is that Jane’s version of Feature X is going to be significantly better than the competitors’ version, simply because it will only have the pieces that customers actually need and use.

The new feature won’t be made needlessly more complicated by bells and whistles that are only put there so that a sales person can check something off on a list. They’ll be put there because they solve a problem.

Sometimes It’s Not the Change They Hate

There has been a lot of discussion in the design world recently about "change aversion." Most of the articles about it seem to be targeting the new Google redesign, but I've certainly seen this same discussion happen at many companies when big changes aren't universally embraced by users.

Change aversion is a real thing. Often people don't like something different just because they're used to the old way. Once they get used to the new way, they discover all the wonderful new features and are happy with the new change.

But sometimes your users' rage isn't change aversion. Sometimes your new design actually sucks.

So, before you blame your users, you should figure out if it's change aversion, or if you really screwed something up. Ask yourself the following questions.

Did You Do Any Sort of User Testing Before Launch?

This is an important one. Sometimes people complain about a product because it has changed. Other times they complain because the product makes them feel stupid or it prevents them from doing what they want to do.

Most often, products make people feel stupid because the products are hard to use.

It's very possible that the changes you made to your product have made common tasks that the user is used to performing harder to do. Yes, the user may eventually learn to perform the tasks the new way, but that new way may be legitimately more difficult! You may even be reducing the amount of time the user spends performing that task, if you make it hard enough.

To pile onto the Gmail redesign for a moment, I can tell you right now that I am constantly hitting the folder button instead of the label button now that they are icons rather than text. I am still occasionally doing this probably a month after I switched to the new look. It's not a deal breaker for me, but it annoys me every time it happens. The new icons, for me, are honestly harder to use than the old buttons, and they build up a certain amount of unhappiness every time I use them incorrectly.

The interesting thing is that this is exactly the kind of problem that you can surface very easily by simply doing some observational testing of people using prototypes or early versions of the product.

Hell, you could probably even figure that one out with metrics. How often are people hitting the Undo button now compared to previously? If people are undoing their actions more frequently, you can bet that your new design is causing them to make mistakes.

Did You Test with Current Users or Just New Ones?

When you have millions of users, it's not cool just to test on people who have never seen your product before.

New user testing gives you really valuable feedback, but it's just one kind of feedback. It doesn't give you any insight into how your current users (often users who are paying you money) are using your product right now.

It may be that your users are doing something surprising with your product. They may be using it in ways you never anticipated. Making major changes without understanding their work styles can destroy something they were relying on.

It's not a matter of their relearning how to do a task in the new interface. You may literally have removed functionality for people who were using your product in innovative ways.

Similarly, if you're only testing with internal users (that is, users internal to your organization), you're not really getting the full idea of how all sorts of different people are interacting with your product. The more types of people you can observe, the clearer your understanding of real use cases and behaviors will be.

Did You Add Something Useful to Users? Really?

Sorry, improving your brand is not particularly useful to users. Even a nice new visual design tends not to be a big enough improvement if you're also changing functionality that users rely on.

Here are some significant improvements that might be enough to counteract a tendency to change aversion:

• making your product noticeably faster or more responsive
• adding a great new feature that users can understand and start enjoying immediately
• fixing several major bugs that people have been complaining about

That's pretty much it.

The problem here is that often companies mistake doing something that is good for the company with something that is good for the user. That can be a tricky thing to spot, but a good way to handle it is to always ask yourself, "What real user problem is this change solving?" If the answer is, "How to give us more money" or "Well, there's more visual breathing space," you might want to brace yourself for the inevitable shitstorm when you launch that change.

Do You Mind Losing a Portion of Your Users?

This is a 100% legitimate question to ask. Sometimes you make changes to your product that you know will piss off a certain subset of your users, and that can be ok.

I've often advocated for prioritizing changes that help your paying customers over changes that help your non-paying users. But there can be other reasons to make changes that annoy certain groups of your users.

The thing is, you have to know that you're making this tradeoff and be ok with it. If you've gone into the change knowing that you might lose a certain percentage of your users, but hoping that you will make up the loss by making other users very happy or attracting new users, that's a fine choice to make. Just be sure that your metrics show this actually happening.

Have You Honestly Listened to Your Users' Complaints?

Let me give you two common examples of user feedback:

• I hate it! It's terrible! I want the old way back!
• I'm constantly hitting the folder button when I want to hit the label button, and I find it really hard to tell which emails are more important any longer. Also, every time I try to Reply to All in the middle of an email thread, it's just ridiculously difficult to do.

Can you spot why they are so different? That's right, one of them is completely non-actionable. There is nothing you can really react to with the first one. You can't fix this user's problem. Yet.

The second one is significantly better because you're starting to get at WHY the user hates the change. You know that the user is having trouble performing specific tasks. You can follow up with the user and have her show you the things that you have made harder to do. You can then figure out if those are things that are done frequently enough and by enough users to justify making them easier to do.

Here's the trick. You can turn the first type of feedback into the second type of feedback by following up with users and asking them for specific things that they hate about your change. If they just keep saying, "It's different!" then they may get over it when they get used to it. But a significant portion of them probably have specific complaints, and writing those complaints off as change aversion is really kind of a dick move.

Have You "Fixed" the Problem by Letting Users Change Settings?

Stop it. Seriously. Just stop it.

The vast majority of your users don't know how to change the default settings.

It's not a failing. They're not stupid. They just don't know nearly as much about your product as you do, so they don't have great understanding of the million different ways you've allowed them to customize their experience. They probably don't even know that those settings are there, and even if they do, why are you making them work that hard?

If you are going to include a few settings that they can change, make them obvious, easy to understand, and don't bury them in a thousand other settings that are incredibly confusing to everyone who isn't in tech and half of us who are.

Like the post? Follow me on Twitter!

I Don't Know What's Wrong with Your Product

When I’m talking with startups, they frequently ask me all sorts of questions. I imagine that they’re probably really disappointed when I respond with a shrug.

You see, frequently they’re asking entirely the wrong question. And, more importantly, they’re asking the wrong person.

It is an unfortunate fact that many startups talk to people like me (or their investors or their advisors or “industry experts”) instead of talking to their users.

Now, obviously, if they just asked the users the sorts of questions they ask me, the users couldn’t answer them directly either. This is the wrong question part. But the fact is, if they were to ask the right questions, they’d have a much better chance of getting the answers from their users.

Let’s take a look at a few of the most common sorts of questions I get about UX and how we might get the answers directly from users.

What’s Wrong With My Product?

I often get people who just want “UX advice.” I suppose they’re looking for somebody to come in and say something like, “oh, you need to change your navigation options,” or “if only you made all of your buttons green.”

Regardless of what they’d like to hear, what they typically hear is, “I have no idea.” That’s quickly followed by the question, “Which of your metrics is not where you want it to be?” If they can answer that question, they are light years beyond most startups.

You see, the first step to figuring out what’s wrong with your product is to figure out, from a business perspective, some realistic goals for where you’d like your product to be right now. Obviously, “We’d like every person on the planet to pay us $100/month” is probably not a realistic goal for a three month old venture, but hey, aim high.

Once you know what you want your key metrics to be, you need to look at which of them aren’t meeting your expectations. Are you not acquiring new customers fast enough? Are not enough of them sticking around? Are too few of them paying you money? While “all of the above” is probably true, it’s also not actionable. Pick a favorite.

Now that you know which of your key metrics is failing you, you need to conduct the appropriate sort of research to figure out why it’s so low. Note: the appropriate sort of research does not involve sitting around a conference room brainstorming why abstract people you’ve never met might be behaving in a surprising way. The appropriate sort of research is also not asking an expert for generic ways to improve things like acquisition or retention, since these things vary wildly depending your actual product and user base.

The appropriate sort of research depends largely on the sort of metric you want to move and the type of product/user you have. You will, without question, have to interact with current, potential, or former customers of your product. You may need to observe what people are doing. You may need to ask them why they tried your product and never came back. You may need to run usability tests on parts of your interface to see what is confusing people the most.

Feel free to ask people like me for help figuring out what sort of research you need to be doing. That’s the sort of things experts can do pretty effectively.

But if any expert tells you exactly what’s wrong with your product without considering your user base, your market, or your key metrics, either they’re lying to you or your problems are so incredibly obvious that you should have been able to figure them out for yourself.

What Feature Should I Build Next?

Let’s imagine for a moment that you have built a Honda Civic. Good for you! That’s a nice, practical car. Now, let’s imagine that you come to me and ask how you should change your Honda Civic to make more people want to buy it.

Well, I drive a Mini Cooper, so it’s very possible that I’ll tell you that you should make your Civic more adorable and have it handle better on curves. If, on the other hand, you ask somebody who drives a Ford F-150, they’ll probably tell you that you should make it tougher and increase the hauling capacity.

Do you see my point? I can’t tell you what feature you need to build next, because I almost certainly don’t use your product! To be fair, even the people who do use your product or might use your product in the future can’t just tell you what to build next.

What they can tell you is more about their lives. Do they frequently find themselves hauling lots of things around? Do they drive a lot of curvy mountain roads? Do they care about gas mileage? What about their other purchasing choices? Do they tend to buy very expensive luxury items? Do they care more about status or value?

You see, there is no single “right way” to design something. There are thousands of different features you could add to your product, and only the preferences and pains of your current and potential users can help you figure out what is right for you.

Should I Build an App or a Website or Something Else?

Another thing that people ask me a lot is whether they should be building an iPhone app, an iPad app, an Android app, a website, an installed desktop app, or some other thing.

That’s an excellent question...to do a little research on. After all, what platform you choose should have nothing to do with what’s popular or stylish or the most fun to design for. It should be entirely based on what works best for your product and market.

And don’t just go with the stereotypes. Just because it’s for teens doesn’t necessarily mean it’s got to be mobile, although that’s certainly something you should be considering. It matters where the product is most likely to be used and what sort of devices your market is most likely to have now and in the near future. It also depends on the complexity of your product. For example, I personally don’t want Photoshop on my phone, and I don’t want a check-in app on my computer.

Talk to your users and find out what sort of products they use and where they use them.

Are You Noticing a Pattern?

Experts are not oracles. You can’t use outside people as a shortcut to learning about your own product or your users. You need to go to the source for those things.

If you find yourself asking somebody for advice, first ask yourself if you’re asking the right question, and then ask yourself if you’re asking the right person.

And if anybody ever tells you definitively what you need to change about your product without first asking what your business goals are, who your users are, and what their needs are, you can bet that they’re probably wrong.

Hey, you got to the end! Now you should follow me on Twitter.

Fucking Ship It Already: Just Not to Everyone At Once

There is a pretty common fear that people have. They’re concerned that if they ship something that isn’t ready, they’ll get hammered and lose all their customers. Startups who have spent many painstaking months acquiring a small group of loyal customers are hesitant to lose those customers by shipping something bad.

I get it. It’s scary. Sorry, cupcake. Do it anyway.

First, your early adopters tend to be much more forgiving of a few misfires. They’re used to it. They’re early adopters. Yours is likely not the first product they’ve adopted early. If you’re feeling uncomfortable, go to the Way Back Machine and look at some first versions of products you use every day. When your eyes stop bleeding, come back and finish this post. I’ll wait.

Still nervous? That’s ok. The lucky thing is that you don’t have to ship your ridiculous first draft of a feature to absolutely everybody at once. Let’s look at a few strategies you can use to reduce the risk.

The Interactive Mockup

A prototype is the lowest risk way you can get your big change, new feature, or possible pivot in front of real users without ruining your existing product. And you’d be surprised at how often it helps you find easy to fix problems before you ever write a line of “real code.”

If you don’t want to build an entire interactive prototype, trying showing mockups, sketches, or wireframes of what you’re considering. The trick is that you have to show it to your real, current users.

Get on a screen share with some users and let them poke around the prototype. Whatever you do, never tell them why you made the changes or what the feature is supposed to be for or how awesome it is. You want the experience to be as close as possible to what it would be if you just released the feature into the wild and let the users discover it for themselves.

If your product involves any sort of user generated content, taking the time to include some of the tester’s own content can be extremely helpful. For example, if it’s a marketplace where you can buy and sell handmade stuff, having the user’s own items can make a mockup seem more familiar and orient the user more quickly.

Of course, if there’s sensitive financial data or anything private, make sure to get the user’s permission BEFORE you include that info in their interactive prototype. Otherwise, it’s just creepy.

The Opt In

Another method that works well is the Opt In. While early adopters tend to be somewhat forgiving of changes or new features, people who opt in to those changes are even more so.

Allowing people to opt in to new features means that you have a whole group of people who are not only accepting of change but actively seeking it out. That’s great for getting very early feedback while avoiding the occasional freakout from the small percentage of people who just enjoy screaming, “Things were better before!”

Here’s a fun thing you can learn from your opt in group: If people who explicitly ask to see your new feature hate your new feature, your new feature probably sucks.

The Opt Out

Of course, you don’t only want to test your new features or changes with people who are excited about change. You also want to test them with people who hate change, since they’re the ones who are going to scream loudest.

Once you’re pretty sure that your feature doesn’t suck, you can share it with more people. Just make sure to let them go back to the old way, and then measure the number of people who actually do switch back.

Is it a very vocal 1% that is voting with their opt out? You’re probably ok. Is half of your user base switching back in disgust? You may not have nailed that whole “making it not suck” thing.

The n% Rollout

Even with an opt out, if you’ve got a big enough user base, you can still limit the percentage of users who see the change. In fact, you really should be split testing this thing 50/50, but if you want to start with just 10% to make sure that you don’t have any major surprises, that’s a totally reasonable thing to do.

When you roll a new feature out to a small percentage of your users, just make sure that you know what sorts of things you’re looking for. This is a great strategy for seeing if your servers are going to keel over, for example.

It’s also nice for seeing if that small, randomly selected cohort behaves any differently from the group that doesn’t have the new feature. Is that cohort more likely to make a purchase? Are they more likely to set fire to their computers and swear never to use your terrible product ever again? These are both good things to know.

Do remember, however, that people on the internet talk about things. Kind of a lot. If you have any way at all for your users to be in contact with one another, people will find out that their friends are seeing something different. This can work for or against you. Just figure out who’s whining the loudest about being jealous of the other group, and you’ll know whether to continue the rollout. What you want to hear is, “Why don’t I have New New New New New Thing, yet?” and not “Just be thankful that they haven’t forced the hideous abomination on you. Then you will have to set your computer on fire.”

The New User Rollout

Of course, if you’re absolutely terrified of your current user base (and you’d be surprised at how many startups seem to be), you can always release the change only to new users.

This is nice, because you get two completely fresh cohorts where the only difference is whether or not they’ve seen the change. It’s a great way to do A/B testing.

On the other hand, if it’s something that’s supposed to improve things for retained users or users with lots of data, it can take a really long time to get enough information from this. After all, you need those new cohorts to turn into retained users before seeing any actual results, and that can take months.

Also, whether or not new users love your changes doesn’t always predict whether your old users will complain. Your power users may have invested a lot of time and energy into setting up your product just the way they want it, and making major changes that are better for new folks doesn’t always make them very happy.

In the end, you need to make the decision whether you’ll have enough happy new users to offset the possibly angry old ones. But you’ll probably need to make that decision about a million more times in the life of your startup, so get used to it.

So, are you ready to fucking ship it, already? Yes. Yes, you are. Just don't ship it to everybody all it once.

Now, follow me on Twitter.