Don't Waste Time on User Onboarding | Hacker News Reader

Don't Waste Time on User Onboarding

greaterdanorequalto.com

208 pointsDanHulton112 comments

I get that there are probably established best practices that people follow for designing onboarding but I've never experienced one as a user that I didn't want to immediately dismiss or skip through.

Likewise. However, it’s worth remembering that the average user of a website that’s called “Hacker News” is almost certainly more tech-savvy than the average user of any app we create.

My solution is, just as the article suggests:

> Okay, so your service is immensely easier to navigate since you ran it past a battery of theoretical users. Now cut half of it out.

Perhaps that applies no matter how big the app gets, perhaps a simple UX that hides the power tools until they’re called for is the one true way. On the other hand, virtually every time I’ve faced a user saying “$App should perform $Task”, it already does. Sometimes the app doesn’t have a particular feature, but the virtual assistant of the OS does.

What’s the best way of helping those users discover those features?

It's more for the "onboarding" session that you'd cut half the app out, I could have made that more clear. The idea is that you're looking to sprint to value, to get the user's eyeballs on the most-important part of your service ASAP. Otherwise, they don't stick around to explore and find all the power tools. In normal usage, you should definitely have an easy way to find and use all features, but that's a broader product design question, not an onboarding one.

Here are our stats for Polar....

Before onboarding our cohort retention was 2%...after 1 week.

This means that after a week only 2% of our users came back to use the app.

After a basic onboarding we jumped to 10%. We're now at about 14-15% (just via improving the app).

About 50% of our users completely skip the onboarding altogether.

I think I'm going to experiment with a video instead of a tour to see the difference.

> About 50% of our users completely skip the onboarding altogether.

How is this different than when you had no onboarding? (serious question) Did you make other changes that made the onboarding process less important to end users?

The glass is half full, not half empty. About 50% of users DON'T skip onboarding - and for those users, it really matters.

You separate people in those willing to explore and figure it out on their own, and those who'd otherwise probably never check further than the home page on their own. In my experience the biggest gain from intros is that when you get people into entering some basic data - creating an example project or something - they feel like they've already invested some time and are much more willing to explore the features further and enter more data and eventually really start using the app.

This is why I have a certain distaste for articles like this. The author has this sort hypothesis / contention, specifically that user onboarding is a waste of time, yet presents no data to either validate or disprove their hypothesis. We're supposed to just accept their hot take as gospel, where as folks like yourself can offer actual data that serves as a compelling counterpoint.

I dunno that it's a counterpoint. The thesis of the article is that setting up a user onboarding flow before you're ready is a waste of time, and/or that it's not as useful an activity as others you could be doing with that same time.

I'd consider Polar a success of "adding onboarding when ready." I just don't want people to consider onboarding as the first and only tool to reach for when trying to fix your Trial-to-Paid Conversion issues. There are a lot of other useful tasks you could - and frequently should - be working on as well.

God I hate user onboarding... seeing stats like that make me so sad, and feel ever further away from the current internet zeitgeist.

Half of their users skip the on-boarding. Why would that make you feel far away from the current internet zeitgeist?

Presumably the user group you've self selected into also has a retention rate that is 20x lower than those that don't skip the on-boarding.

Because everything has onboarding now. Sometimes it's unskippable (think mobile games). It's always annoying, very frequently patronizing, and, frankly most people's app/site designs aren't that special because modern UX seems to have no more than a handful of designs for how tech is to be spoonfed to the masses.

Of course none of that matters because it makes another tiny, self-selected group considerably more money.

This is really interesting, because Polar is the first onboarding experience I've had in a really long time that made sense and wasn't extremely annoying.

What is the retention of people who do vs don't skip the onboarding? Is it a wide disparity?

can anyone else share some A/B stats?

Same. 9 times out of 10 I don't want to learn a new tool, just use an intuitive, predictable UI. If I need to "read the instructions" first, your app is probably trying to do too much.

Keep it simple and only use onboarding if your app is truly novel.

If a product wants me to go through a lengthy onboarding session, whether it be with a representative or through some guided tutorial it's an immediate turnoff and I will likely abandon right then and there.

Personal anecdote:

We were evaluating Cisco AppDynamics last year as we were in the market for an APM product and, unfortunately, we are a huge Cisco shop. They wanted to hold our hand and walk us through everything, and the product was certainly in a state where I can't blame them - the agent sucked to deploy, they had a nice little UI to configure things during install but after that it was all editing XML. Sure, I could automate it all with Ansible - but you could tell they knew the UX sucked and they just wanted you to get through the onboarding quick and hope you didn't notice.

We ended up purchasing Dynatrace instead, I started a trial with no intervention from sales reps. They made their happy onboarding path a "game" (they have "quests" in the UI to guide you through the various parts of the product) but otherwise threw you right in with no handholding. The agent was stupid simple to deploy (here's a manifest to deploy to OpenShift, here's a .msi to install on Windows) and it picked everything up automatically and started working. Figured everything else out by just playing with the tool, with the docs as an aide.

// End anecdote

I think onboarding has a place, but it shouldn't hold a user hostage. I really liked Dynatrace's approach, especially as the "quests" were covering different components of the product and could be tackled in any order - they also weren't guided tours but instead directed me to investigate other parts of the tool and perform certain actions, even providing a carrot (free credit) for completing them.

This is a problem I've run into anytime I'm purchasing "Enterprise" licenses for software. They always want you to contact a sales rep, talk to an integrator, set up a meeting to discuss your needs. My NEED is to pay via Purchase Order...

There was never an enterprise licensed software that didn't involve multiple meetings. The software was the same software we had already integrated in the development environment, we just needed licenses for production. Didn't matter that we'd been using the software for ages, they still wanted to set up hours of meetings before sending us an invoice.

If you can skip it, it’s probably not designed right. Well designed onboarding should answer the questions you might have right when you have them, not before you even know that there is a question to ask. It’s not rocket science, but it takes some thought. Design empty states that make it obvious how to use a screen and what it’s for, always have a clear “main” action that will have a productive outcome for most users most of the time, and make sure to have a clear and consistent way to get well designed help that is relevant to what you currently see on screen (can be tricky to do on small screens, though).

That said, traditional onboardings have their place, for example to introduce new features. Even if you skip them or click through and only skim it, you might still pick up enough of a heading or illustration to make it easier to figure out why things are different.

I have!

Exactly one, actually..

RoomieMatch.com. Shoutout if you're reading, dude/dudette.

Something I've noticed is a lot of companies use user onboarding and educational messages as a crutch for confusing UX and product experiences.

Many companies also have a one-size-fits-all approach to onboarding, where they send the same messages to all users, regardless of their in-product actions. Instead, it's better to trigger relevant messages if the user has taken a specific action and hasn't taken the next step to achieve Y result.

You can always send onboarding sequence to users who haven't taken any actions at all.

It should also be noted that for some established companies, a lot of their onboarding isn't necessarily to educate you but to force you to take certain actions that their testing has shown will make you active. For example, Pinterest forces you to follow certain topics when you sign up.

He has one good point about user testing. The way to test a user interface is well understood, but not done enough. You have a setup to take video of the screen and the user at the same time. You get some people roughly comparable to the desired user base to try it. Alone, with no help. You give them a set task to accomplish.

Then you have someone review and annotate the videos. Where did people get stuck? Where did they back up? Where did they give up? Put together a highlights reel and have the developers and managers watch it. It's usually obvious by that point what's wrong.

My only point of contention after skimming through the article is with the trial period. The other side of that coin is that, if it's too short then users bookmark your site thinking they'll sign up later and forget, and if it doesn't exist then users may end up not signing up at all and going with the free option instead. In practice it really depends on the product's criticality.

Came to write this. Companies forget that their product isn't evaluated in a vacuum, that for any serious business need, a person or an group is going to evaluate multiple products for a extended period of time. Having no trial period or a very short one means losing out of business simply because of short term greed. Also, it doesn't help that psych pop is used to arbitrarily justify any conversion tactic.

It's an optimization problem - you lose X% of people if your trial is long enough that they don't prioritize evaluating it before it expires. You lose Y% of people if your trial is too short for them to evaluate it against competitors. If X > Y, shorten your trial. If Y > X, lengthen it.

The problem comes in the fact that it's really hard to identify what X and Y are - there's some voodoo involved in properly attributing actual reasons for not converting a trial.

This is why I like the idea of going immediately paid with a very-long money-back guarantee, though that comes with its own optimization problems.

In my mind, there's more to it than a metrics-driven optimization in isolation from the rest of the company.

In your example, the Y% of people you might lose if your trial is too short may very well turn out to be higher paying and/or lower churn customers. What you suggest in your post is the marketing equivalent of sales and support ignoring the colloquial small fry, without realizing that the latter might be sitting on company boards and referring large clients.

Also, if they bookmark the service until they can set aside a time to sign up and evaluate (which as an occasional buyer of SaaS products, I can anecdotally report that some do), you'd have no way to reach out to them and build a relationship unless they contacted you.

By contrast, if you have a free tier, and/or make it clear from the outset that when your trial expires you'll be able to do new ones when the product hits major milestones, then you have the end-user's email, you're building a relationship, you can interview the end-user to understand why they didn't convert, you can try to rescue trials that didn't move forward with a sale for a reason or another, etc.

This article is great, and full of really helpful advice!

I think the best advice for me was "Add sample data to avoid an empty nest." I just realized that my product has this problem, so I'm going to be adding some sample data for new users. I think I'll also have a toggle switch so you can easily hide all the samples. (It would be annoying if you had manually delete everything.)

Concierge setup is also a great idea, and I've recently been offering that to new customers.

> I think the best advice for me was "Add sample data to avoid an empty nest."

I respectfully disagree with that view and I say that as someone who has implemented a product with sample data during the onboarding process.

Users don't care about your data. They care about their data. The easier you make it for the user to input their own data and get some value out of it...even if its something trivial...the faster they'll be able to relate their problem to your products ability to solve it.

I've been thankful of sample data before - eg auditioning online image editors, I didn't want to hunt around and upload something but just wanted to get right in to editing to get a feel for the apps and whether they're so what I needed.

In general though: most onboarding I've disliked, too slow, but some with "click here to do this, or elsewhere to end the tutorial" have been good.

I don't know that we have much to disagree about. Absolutely users will get more value from seeing their data, but there are situations where it can take a while to get "their data" visible, like if you're waiting for people to answer a form that they created with your software. Instead of having them wait a day, it can help prove the value of your service to show them what data will look like, once it starts coming in.

Instead of adding sample data when the initial view would be empty, I prefer to replace the empty view with a message about what the user should be doing first.

So instead of "0 photos" and a blank page, it could say "Get started by [uploading some photos]".

This is great UX advice! Turn an otherwise empty view into an invitation.

It annoys me when onboarding basically means having to delete the demo data that was automatically added because I like my apps clean and tidy, and because random data makes it difficult to discover the app's functionality my myself. Maybe there's a compromise where you can suggest to the user that they can import demo data to see the potential of the app?

When I tried Asana, they had lots of example projects you could choose to add. These showed what's possible with the more advanced features.

I really like this approach because the use case for products like these varies from one client to another. Just throwing in some demo data that's always the same wouldn't really help everyone.

Sample data can be helpful, but in my experience, usually isn't.

What users actually want is a quick way to work with their own data.

A specialized "first launch wizard" that makes entering inital data really quick and painless is often a better choice.

It can also make sense to allow trying the product out with existing data, but for this I would offer a "demo" mode where you get bootstrap data, can play around with it, and have a big banner on top that allows launching the real thing.

This depends on the product of course, but I prefer those two options above.

The article talks about “user onboarding” as if it is different from the first run experience. Perhaps it is a SaaSy jargon to separate them. But for us at Sleek, we obsess over the first run as it is how we onboard users - until the user has gone through their use case, we assume they’re new to it.

Learn by doing, not by learning.

It's one of those weird terms that seems to mean different things to different people. For some, it means "first-run experience", and for some it means "the tooltip hoops you make people jump through". I think for too many, it's the latter.

But yeah, many people only get a first run. You have to obsess over it, because if that doesn't work for you, it can be like pulling teeth to get people to come back.

On the other hand, if you're offering something that breaks with an established paradigm -- i.e. first-time viewers don't know what to expect, a short storyboard to clarify your model might be invaluable, and worth fussing over.

This is the "don't waste time" part - on-boarding users with things they already know is wasting time, so don't do that. Teach users the new and exciting bits they don't know yet.

I'd venture to say that breaking UI paradigms shouldn't be taken lightly. Your application is not special to me, the user. I don't want to have to learn how to use the interface of dozens of different applications because they all thought they were special snowflakes and could break established paradigms.

I was referring to application paradigms, not common UI patterns.

What's the distinction? If a user is surprised by what happens when they interact with the UI, the UI has failed.

If we're talking about an application that requires a lot of outside knowledge (e.g. GIS (Geographic Information Systems (mapping)), accounting, video editing), you're not going to be able to "teach" the user what they need to know in the UI or some short videos.

In one case on-boarding is a symptom of a bad UI, in the other it's more-or-less useless.

But how do I get visitors in the first place, this is the first challenge. Twitter is pretty good at hiding my promotion tweets from other users...

No idea if this is common practise, but I block any account that promotes a tweet in to my timeline so I don't see it again. If it is, and you're serving a niche market with few people who'll buy from you, promoting tweets could be a really bad idea.

So how do people reach you to sell to you? Or, to turn it around, how do you find out about new products/solutions?

I don't use Twitter, but if I'm specifically looking for something, I will actively seek out and do the research. I suspect many others who block ads and such are the same.

"They don't reach me, I reach them."

What about when you are not specifically looking for something? How do you find out about new shirts, or cars, or coffees or trips or whatever? Do you stumble across them on web sites, like in articles? On TV? If you usually find out things word-of-mouth, then how do your influencers find out about them?

I'm not the OP, but I'll answer the question anyway. When I'm not specifically looking for something, then I don't find out about them. My attention is limited, I don't have enough to give to every little outfit interested in taking my money. I don't need a new shirt, or car, or coffee or trip. And when I do, then I'll google it and I'll happily see all the ads that Google will show me.

The ads are blocked until then.

OK- so ads, then. At your convenience, to be sure, but ads nonetheless.

I am experiencing an almost 100% disconnect with almost everything you wrote.

I don't find out about them then, and my savings and 401K continue to grow unabated. Maybe I'm missing out on cool stuff, but assuming I never found out I was missing out, then my life isn't impacted negatively regardless. I don't really know what an "influencer" is; is it the internet equivalent to the ShamWow or OxiClean guys?

I have 0 tolerance for people or companies actively telling me what I should buy or like, and distrust anyone who does so.

I'm going to present an alternate take on "how to live" from you and several other people who replied, not to say your way is wrong, but to say that others live other ways.

I like hearing about cool stuff, I like sharing it with friends. I like style, looking fashionable, and being apprised of trends. I don't plan to die with savings.

You wouldn't trust me, because I often tell my friends about things they should know about; not because I have anything to gain, but because we all do this in all of my social circles. If I hear about a nootropic they might like, I tell them. If they know of a bar I should try, they tell me.

I like a lot of ads because I like knowing what's going on in the world. Even in a post-capitalist Star Trek world, I'd want people who made stuff to tell me.

Neither way of living is wrong.

Don't take me the wrong way: I hate ads too. I usually find out about things from other people, or ads in non-online media, or when I am physically at a store or on Amazon searching for something.

I tend to by generic/commodity items over brands, but there are some notable exceptions.

I pick my groceries by checking the grocery ads and seeing what's on sale.

I don't think I've ever clicked on a search ad or a mobile ad except accidentally.

Concerts and stuff I learn about through ads.

It's kind of an eclectic list, I guess, but ads do seem to serve some purpose. I'm just hard to reach via them unless I'm going looking for something.

I don’t want to find out about any of those things. They don’t, as the saying goes, spark joy in my life. Sure, trips might be the exception from your list, but also any trips I go on I can plan for myself in a way that is both cheaper and more enjoyable.

That’s why paid ads on google search exist; so people can reach them easier

That's what Twitter is really good at already, it's literally just people chatting and shouting about stuff that other people are chatting and shouting about.

I can't imagine Twitter being good at reaching a market unless you've gone viral. Maybe I'm wrong, but it seems like it's like pouring a taste of wine into a fire hose. I don't think I've ever learned about a product on Twitter (but I block ads and don't do much on there). I hear about news and stuff from Twitter.

You have to have a following and be careful with how your market your stuff but yes, I've discovered many products via Twitter BUT they are almost always aligned with my interests and hobbies. Those that aren't, typically promoted tweet, get blocked.

Yep, same. And it's word of mouth from people/entities I trust. And that trust has been built up over time. I have zero attachment to Acmecorp or Joe Random, and generally zero trust in them.

I've heard others do this as well and I don't quite understand it. If you don't want to see ads, sure maybe use AdBlock or PiHole, but why block the company?

Not trying to put words in your mouth, but are you offended that the company spent money on a promoted tweet? That it was an overreach on their part?

If you don't want to see ads, sure maybe use AdBlock or PiHole, but why block the company?

Blocking the company stops the tweet appearing everywhere I see tweets. That's on my home computer, laptop, Chromebook, office computer, my phone, etc. It continues to block the ads from that company if/when Twitter work out ways around ad-blockers. And it makes me feel like I'm proactively looking after my account.

I use ad blockers as well.

are you offended that the company spent money on a promoted tweet?

I just find that Twitter promotes the same tweets over and over again, so if you don't block them it gets very repetitive. If I was actually offended I'd reply first so they had to pay for an engagement as well as the promotion.

I don't want to see ads, ever. I hope a growing number of people feel the same way.

I don't care about the business model of the whole internet. It is obviously a house of cards. A business model based on bad UX? I hope it collapses.

No, I don't know what will replace it. I don't care, as long as it doesn't involve ads.

I really don't care if the ads are innocuous or labeled or relevant. I don't ever want to see an unsolicited nudge to spend money on something. Ever. I can't be any more clear than this. I hope there are more like me every day.

We're out here. You're not alone.

I completely agree.

I don't care if your ad is clever, relevant, organically sourced, environmentally friendly, necessary for financing your product, something I actually want, need, or didn't know I wanted yet.

The internet is plagued with this cancer everywhere you go, no matter what you do. Somehow it's become acceptable that every second of our attention and time is competed for by corporations trying to sell us things.

I simply do not want to see ads, ever.

I don't understand this overreaction. Ads are there, so what, they don't even make noise or tend to cover parts of the screen like they used to.

The point is that I'm fundamentally disgusted by an incitement to spend money being thrust in front of me in any form, for any reason.

I see it as exactly at odds with what I'm trying to do with my life and my attention. I do not consent to have my time and attention taken from me in this way.

Yes, I am a free rider on much of the internet today, such as it is. That the entire internet is based on this absurd proposition is a symptom of its rapid growth and industries going after the lowest hanging fruit. I am confident that the internet will find another way forward.

You need to look at what powers the ads.

It's never "just" an ad. Like the way you see an ad in a newspaper. It's the spyware that comes with it and the fact that tons of them are animated!

And yes, I consider it spyware and virus-like behaviour to track me wherever I go and I will actively do what I can to stop it. Even if it means you don't get paid... you chose that seedy medium to make money, not me. It's not my job to validate your business model.

By all means block me. I will go elsewhere.

Browsing the internet without an ad blocker is actually impossible for me now... whenever I rebuild a machine I have those few hours or so before I get the ad blocker installed where the internet is a horrific assault on the senses with everything trying to steal my attention away from the thing I am there to see and slowing down the page load (significantly in some cases)

I still remember when plenty of the internet was running fine from hobbyists and whatnot without resorting to ads and tracking... it costs practically zero to run a website (granted, at scale it's a whole other ball-game but how many people have that problem?)

The internet would survive fine without ads. It would have a different mix of behemoths and hobbyists sites but it would survive just fine.

That being said... I do remember trying to punch the monkey to win :D

Edit: added stuff for clarity

The ones under discussion are the digital equivalent of junk mail

To jimmaswell:

Ads wouldn't be there unless they work for the advertiser.

Unless the advertiser mistakenly believes they work.

I used to feel that way, until I built several services (the latest being tomotcha.com). Turns out, once you've built something, you need to advertise its existence out there. I mean "advertise" in the largest meaning of the word, not only paid for advertisement on Google/Facebook/Twitter/Criteo. That includes writing blog posts for people to stumble upon, talk about it on Hacker News, send samples to reviewers, etc.

How else would people learn about it? Turns out, if you don't talk about it, very few will come...

I think this is why HN has a lot more goodwill towards self promotion (Show HN style posts) than a lot of other communities (reddit notably, depending on the sub). Entrepreneurs know that advertisement is a necessary part of any business.

I think people confuse advertising in general with the 360 tracking facebook/google and others use. People need to advertise and it's helpful to both the consumer/company.

I'm not sure that people confused the two, it's just that the intrusive forms of advertising make up an overwhelmingly large part of a normal person's interactions with ads that we all just assume that that is what you're talking about.

I think we mostly appreciate the need for the so-called "getting the word out there", but it ought to be done in way that's appropriate. There's a balance to be made between the potential value of a product, and the amount of attention you're trying to steal. Most people think the products they make are more valuable than they really are.

Vivaldi (Chromium-based browser, built for power users) has a set of default home page spots that companies paid to get a spot on. They are removable, and are a sorta one-time thing. They help pay for development, and are rather non-intrusive. Is that acceptable?

What about a set of sponsor logos at the bottom of a webpage that let you visit the companies?

The take-away here is to just buy covert ads that you confuse for content. It's win-win. You get to think you're not seeing ads, and advertisers get to advertise in unblockable ways.

I don't really have a problem with that. If you're making content, and it's valuable, and it's related to a product, that's great. If your content isn't valuable to me, then I won't bother with it.

If you're just getting folks on facebook to promote your products in the hope of some minor dispensation or something (a popular gimic these days). That's easily identifiable and blockable.

The rest of us don't want to pay a monthly subscription to every single website we use, so ads it is.

Ad supported also competes with entirely free, by having more resources for development and, go figure, more marketing/ad dollars to promote their services elsewhere and gain usershare, for cases in which everyone-is-using-it matters—so, social media and messaging. Paid services aren't the only things that can/would replace spyvertising-funded services.

One grey area that I approve of is referral links, if they're obviously presented as such.

For example if you read a tutorial on how to build a weather station and the author provides referral links to a component on Amazon, I'm happy to use those links.

Referral systems are open to abuse though, like the myriad of ultra specific comparison websites with tenuous "expert" reviews.

I agree with this. An ad that is relevant and doesn't track, harass, or deceive users is reasonable in my opinion.

There's no reason that most ads couldn't fit this model, but due to greed and laziness they mostly don't.

@rurp-

A lot of those referral links do have logging (seeing who visited / hoe many people have) and tracking (general location, some demographics). Is that okay to you?

In my case, I guess I am? Obviously the site owner will get some logging results, and maybe they will try to optimise for that.

I'm fine with a site owner knowing their demographic. That's potentially useful for other things, eg should you invest in localisation? Otherwise unless you have a vpn, your IP is going to be visible anyway.

It probably depends on the referral site. For example, do Amazon have any reason to sell my data to anyone else? It's most valuable to them and they already have a decade+ of purchase and browsing history.

I still think this a preferable situation to targeted ads which try to guess what I want.

Are you saying you want something for nothing? If not nothing, what?

Ads aren't just the business model of the internet. They've been subsidizing human interaction since ancient times.

However, they weren't always backed by vast networks designed to stalk, and manipulate you, with access to more data than most ruling governments have.

Have you ever ran a business? Curious to how you promote a business without ads or sponsored content to get the word out there to the masses..

On that note, I have nothing against sponsored content. For example, a blog on programming that happens to have posts sponsored by a company that specializes in dev tools or something.

That makes perfect sense!

Sure, some of it can be puff-pieces for the company and they are quite clearly just bullshit but it's easy to leave and go elsewhere if they take the piss.

A banner at the top of a blog post saying "this content was kindly sponsored by <insert company relevant to the content>" is a great idea and more should do it.

But they don't!

They use ad networks that track the ever-living shit out of every move you make.

That's the issue for me!

Good old SEO?

I mean if your product fills a unique niche it should be really easy to rank high. Get your first few customers like that, then because your product is revolutionary, word of mouth will do the rest.

A real-world example: Monzo, a modern bank in the UK. They only very recently started using advertising, but got their first million customers with nothing but word of mouth. Their secret? Simply making a product that's light-years ahead of what the existing banks provide.

Your product is not revolutionary? Oh then I don't know (actually I do - make a revolutionary product - switch industries if necessary), but you're not entitled to waste my time, attention & privacy with ads just so your business can survive. Being profitable is not a right, I couldn't care less if you go out of business.

SEO, good content marketing, sponsoring blogs and videos (when done tastefully), product placement, engaging with communities, talking about it in HN/ProductHunt/etc.

All those things are acceptable to most people who block ads, IMO.

Some of those things still sound like ads. I’m not convinced that a company, for example one that sells shoes, would not profit from running ads everywhere

Promoted tweets are generally annoying, I can't speak for others but it's just slightly satisfying to tell the advertiser to eff off for breaking my reading flow. I literally don't care that the company is trying to sell me stuff, it's a company, why should I? I follow people, companies and orgs that I'm interested in, and Twitter suggests similar and I often take them up on the suggestions. Those users will advertise if they're selling things in tweets, I don't need Twitter spamming me with adverts as well.

Not a Twitter user, but isn’t that just the same as ad blocking? Company X pays money to insert their message into person Y’s life: that’s an ad.

If someone doesn’t want ads in their life and Twitter provides a tool to block them, why wouldn’t that person use the tool?

A lucky HN or PH campaign might help for a while, but the traffic spike will calm down soon.

I'll pay for them on the long run. For me, the CPC cost will make it worth.

There are also other ideas, like launching an affiliate program, app marketplaces.

Look at where the traffic is in your niche.

I was a bit sceptical when I read the title, but the author raises a good point about calcifying users into suboptimal ux decisions.

Wofkflow breaking UI changes are always hard and there's always someone whose use case [you're breaking](https://xkcd.com/1172). Then again, you can't please everyone and it might hurt to stick with older decisions.

Onboarding itself is extremely valuable. No application is so self explanatory that everyone gets it without a guide. When should you onboard then? I'd say when you're sure that you're not doing any major changes within the next few months.

All UIs can and should be iterated upon and it helps if you have a timeline on when you're upgrading which views and components and it'll help you decide when to launch the initial onboarsing process.

> Onboarding itself is extremely valuable.

As a user, I find them extremely annoying. They stop me from achieving what I want to achieve and I try to get rid of them as quickly as possible. I also find them as the hallmark of a bad UI -- if you have to tell me how to use your product, then it's not well designed.

> if you have to tell me how to use your > product, then it's not well designed.

VIM users, licensed drivers, and pilots would disagree. Powerful tools require training. If you're pushing an app that I'm already familiar with before I've ever used it, then you are obviously not innovating.

Perhaps we're talking past each other? To me "training" is different from "onboarding".

"Onboarding" is something that happens quickly and right at the start -- OP links to blogs about building those screens that block the whole UI and point out some buttons. It's something that isn't meant to be all-encompassing. It's something to point out things you wouldn't expect. (That last bit is why I dislike them so much.)

"Training" is something that isn't expected to be fast often continuous from the start. It's meant to be fairly encompassing, as in you should be pretty well versed in something once it's finished -- not an expert, but some experience and understanding of the whole. Training is meant to include things that aren't obvious/you wouldn't expect (e.g. steer into a skid), but is (and I would argue primarily) about building a foundation for the UI and for further training and learning. Training also includes practice.

Yes, vim has a learning curve and breaks what has become the current UI paradigm. However, I'd argue that unless you're just using vim occasionally, you need to be "trained" on it, not "onboarded". That "training" starts with getting the user to understand the UI paradigm, which isn't expected to be quick. By the end of vimtutor, for instance, the modal paradigm and the ability to chain commands together should be clear and practiced.

Driver's licenses take at least 6 months and both a written and practicle test to get initially. Commercial licenses require training from accredited institutions. There is no expectation that this will happen quickly. Driver's training includes practice study (e.g. driving with a licensed person present). Driver's training also included a study of the UI (e.g. signs of similar styling convey similar information, for example green is an information highway sign, white, tall rectangle is a speed limit, and yellow is important such as exit only lane) which is partially tested in the written exam.

Ditto for powerful tools, you're not expected to be told "this unmarked button does this" and be left to your own devices.

These aren't things you could ever "onboard" someone to do. They require actual training.

> If you're pushing an app that I'm already familiar with before I've ever used it, then you are obviously not innovating.

That's very judgemental -- innovation doesn't mean confusing the user. Did Google's original/classic UI mean that it wasn't innovative?

There are applications that I would argue require "training", even if of an informal type. This doesn't mean that someone couldn't poke around in the application and get something done, but they probably won't be able to exploit the application to its fullest. Two example: GIS tools (i.e. require knowledge _about_ GIS, _about_ projects, _about_ UI and spatial concepts) and Video/Photo editing software (i.e. channels, color spaces, types of transforms, &c). Again, you can open QGIS and GIMP and do something meaningful on your first time, but that doesn't mean you won't feel lost or that it'll be easy.

As VIM user, a licensed driver, and a pilot, I would like to point out that there’s a huge difference in risk of death and injury between those.

Learning vim can be done quite safely by trial and error, then googling for how to do specific things when interested. The other two, not so much.

So unless you’re product has a risk of death or severe injury then I will agree that it shouldn’t need training.

About planes. Unlike cars, they don’t have a standard cockpit layout which leads to lots of bad design In my opinion.

Don't waste too much time on user on boarding, but putting some time towards user on boarding is good exercise to see if your product is understandable and are any holes that may off put users.

It has it's place for the right use case. Superhuman requires you to go through a 30 minute on-boarding one on one with a person. Superhuman isn't the most intuitive if you don't take the time to learn it as it's efficiencies come from keyboard shortcuts and features that differ from established email norms. For this kind of B2B app with a certain level complexity a seamless on-boarding makes all the difference. I never would have use the app long term if not for the process.

While the article has done good points about what to do instead of an on-boarding, I'd also suggest on-boarding can be good for the user. Video games tend to do on-boarding, and often it's not skippable and covers very basic idioms (use the left stick/WASD to move).

The difference I think it's the novelty of the new environment and a sense of accomplishment that it rarely is a chore. So go ahead and on-board if the user gets to see something very cool.

If you need to create on-boarding your product is to complicated. Add features over time, measure there impact and remove them if they don’t work.

I disagree. there are products that are complicated, because the underlying problem your product helps to solve is complicated. And if there are existing solutions that tackle this problems differently to you, you might have to give users a pointer.

If you e.g. design a software for electronic circuit design, good user on-boarding will allow you to highlight specifically why your product shines in comparison to other products.

Another objection: some “products” are specifically meant to work for power users and a increased software complexity can increase the working speed and power you give to the user at the cost of making it harder for new users (e.g. vim for text editing).

Of course if your product is the equivalent of a egg timer, it should be 100% self explainatory. And even if you do a electronic circuit design software it is always good to keep things self explainatory where you can.

But sometimes you have to explain things in order to equip your users with a new, innovative and powerful concept they might never heard about before. Removing that concept to avoid complexity could them be akin to changing your target group. You can fl this, but it needs to fit the concept.

Oh and sometimes you can just do both. Give your app an expert mode.

Do you honestly believe you can onboard someone to electronic circuit design software, by having a bunch of annoying modal boxes popping up showing you where some key features are? I agree there is complicated software, but that isn't made less complicated by silly onboarding. And if your software is not complicated, you're just pissing people off having to click away from all that junk. If it's not complicated, but confusing, that's a UI design issue that will not be solved by a 3 minute intro.

So you take the idea of on-boarding in it’s most horrible manifestation and use it as a argument against the fundamental idea of introducing users to new concepts within the software they are using?

Yes, because the vast majority are as I described. Onboarding prompts are like those EU browser notifications multiplied by a hundred. No one cares, no one has ever read them. We are autotuned to just click away.

Is that really the most horrible manifestation or a typical onboarding experience in practice?

Most of the onboarding I've seen clearly has some real thought put into it, but it still feels modal.

They usually involve widgets that pop up and must be dismissed. Often they pretend to not be modal by not covering the screen, but I assume I have to chase them down and make them go away.

The perception of complexity matters more than actual complexity.

At my job we made the product more complex and in rushing to release it, we didn't include an onboarding process like the old one. The new product ended up winning the AB test in conversion.

As a follow up, I did a 3 way AB test that included the original product without the onboarding process.

What I found out was that the original product without the onboarding process performed just as well as the new one. It was the onboarding hurting our conversion.

We originally thought those extra steps helped the user make a decision but it just made the product seem more complicated and lead to analysis paralysis.

Yeah it's the user's perception that is the key.

I work with a product where the end user is ... shockingly luddite. I suppose everyone feels that way but these end users are very very easily intimidated even with little changes.

Accordingly it doesn't matter how much I think something is simple, it matters what they do and what they're intimidated by. I could make the best super cool web app ever, but if they don't use it / effectively it doesn't matter and becomes a headache for everyone.

Don't waste time on writing an article about introducing new members either. just go immediately into another customer manifesto. I appreciate that the customer has to be the focus, but this article was a bit bait and switch, and didn't give any time to the legitimate issues of growing a startup team.

One of my favorite resources for onboarding is Samuel Hulick's https://www.useronboard.com/

His teardowns are particularly valuable to see how others do things and what does and doesn't work. Plus they're entertaining.

Absolutely an inspiration of mine. If anything, my only quibble is that he doesn't go far enough. I'm working on a similar series, except I'm examining the entire Trial-to-Paid flow, from the moment you sign up to the moment you convert, including lifecycle emails, payment flow, etc. Takes a while to do, though, since you usually have a month or more before the email cycle is exhausted.

Strange title for an article analyzing onboarding.

This is terrible advice. And Click bait.

I'm fact the article gives a lists several things you should do to onboard customers.

Reply on news.ycombinator.com