Hacker News new | past | comments | ask | show | jobs | submit login
Why We’re Pivoting from Mobile-first to Web-first (philosophically.com)
125 points by olivercameron on Nov 27, 2012 | hide | past | favorite | 57 comments

A very interesting read indeed.

The question that keeps bugging me is that maybe this guy is blaming the platform, but the real problem was that they built something people didn't really need.

Look at http://any.do, these guys launched mobile first and hit millions of users in an extremely crowded market(todo lists) because they've created a great product a lot of people want and use daily.

In the end, it sounds like(and I don't know their story from close, only judging from things written in this post) they raised too much money early on because they are a YC company(raising a lot of money actually makes you much slower), launched a product with a shitty signup funnel and took a lot of time to fix that.

Apple taking their time to approve new versions isn't new, anyone(especially a YC company surrounded with top mentors) knows that a/b testing in a mobile app is a problem that should be tackled upfront.

Personally, I'm tackling that problem by testing everything I can BEFORE I even start to write code.

I'm running ads on FB/Google, testing CTRs for different value propositions/positioning, driving users to a landing page where I'm testing different mockup ideas I have for each value proposition, testing the icon, testing price, testing feature list.

I'm talking with as many customers in my customer segment I can not only to understand their needs and problems, but also to create an initial user base that likes me enough to agree to install my app BEFORE I launch it in the appstore using TestFlight.

This will enable me to discover a lot of UX problems upfront and to test a very very very important question - Is what I've built a painkiller or a vitamin, which is the most important question (I think) mobile app devs should ask since it's clear that most apps are downloaded and then forgotten in the abyss of unused apps.

If someone has read this long reply and has more ideas on how to test this question before launching, I'd love to collect more ideas.

>The question that keeps bugging me is that maybe this guy is blaming the platform, but the real problem was that they built something people didn't really need.

A thousand times yes. The market is way past saturation. It doesn't need yet another social network that will never ever be as successful as facebook.

Have you considered qualitative surveys and relative importance testing? I've found that useful for validation in the past

This post seems to conflate two issues:

* Churn is high, optimising user flow is mediocre at best, and iteration speed is slow, on mobile.


* Free is bad because of privacy blah blah blah.

I personally view the first issue as kind of un-solveable. Life is hard, etc. What are you going to do, make a HTML5 mobile application? good one.

The second, ehhhhh. Users don't give a shit about privacy. Really. Really. Like, really. Hell: I'd rather use Facebook for free and let them "sell my data" versus pay $50 a year.

Anyway, I'll reiterate my #1 learning of how startups grow into businesses. There are two ways to build a $100M company:

#1 Your users have a high LTV (=$5k+) and you can afford to spend significant (=$100+) capital on acquiring the user.

#2 Your business is massively (=10) viral, allowing you to acquire users insanely cheaply.

The reason why it is – always – very questionable to charge for a B2C user is that it's generally rather difficult to have a story that makes individual people = high LTV, ironically, without having millions of them.

You need to wrap your legs around one of these two approaches and never ever stray from it.

I am making two points, the first that we are web-first, and the second that we are revenue-first. I'm not conflating them; that's the point of the article.

In any case, they are related in my mind. In order to justify why we are switching to a paid model, I bring up privacy because I think you are trading the loss of privacy for that of cash when you choose between ad-supported and a paid model on a consumer app. As for going web-first, I think I made the reason clear.

I agree with most of what you said. Sharing needs to be #1 to have viral growth. Large audiences are required for IAP and Advertising to make sense (I'd also add freemium to that list). An intrinsically leaky funnel needs even heavier viral growth to get to those large audiences. Heavy viral growth is probably going to be at odds with privacy.

But I think you need to pay attention to what others are saying. If you can't get users to complete a series of tasks to get to your value proposition, either the (expected) value you're offering is too low, or the expected probability of getting that value is too low.

In other words, with a high enough value proposition, you should theoretically be able to get someone to do something EXTREMELY difficult.

I agree with your points, however, you don't need anywhere near $5k LTV to be able to spend $100 to acquire a customer. Also, I'm unsure what "10" viral means, but with k>1 virality your business will grow exponentially by definition.

ironically i chose very high example numbers to avoid a bicycle shed debate as to what required numbers are.

I think the points about slow and frustrating iteration on mobile are valid, which is my argument for going mobile second (http://jamiequint.com/why-i-believe-in-mobile-second) However, I'm dubious that the issue with mobile is a more difficult conversion funnel.

The issue is not that mobile screens are too small to show more than one step at a time. I've done many landing page experiences on the web that would have only fit on a mobile screen and seen upwards of 75% conversion (for a 16 question onboarding process) and often find that multiple steps convert much better than a single step (asking for email first, then asking for a password, for example). The key is to get buy-in, which is a combination of value proposition and the give-and-take nature of whatever signup flow you are developing. You can't just ask the user for things over and over, give nothing back and expect 50% conversion. You have to make the process feel like a game, get buy in, ask for a bit more, give value early then ask for more. This is hard if you have a social product because you have to have strong social pressure to signup or a good "single player mode" (http://cdixon.org/2010/06/12/designing-products-for-single-a...) or you are dead.

Great points, your article was spot on. No, definitely that's not the main issue, although it is one. You can't keep reinforcing why the user is signing up throughout the process the same way you can on the web. All mobile sign-up screens pretty much look the same and I see a lot of users look at it and just delete the app, or delay until later at which point they forget why they downloaded it. And even when they get past it, if they see something else they perceive as a "this is going to take a while" moment, they are going to leave and delete your app. On the web, the time to signup and get use from your app is faster and perceived as faster so I think given a choice between the two, more users would do it on the web.

Most landing pages on the web look the same too. The best signup flows I've seen or created don't need to continually re-enforce why the user is signing up, they just need to lead the user down a logical path towards fulfilling the original value proposition, the reason they typed their password into the App Store and took the time to download your app in the first place.

If the user is bouncing because they think "this is going to take a while" then you either don't have a strong enough value proposition or you're making the user think too hard. Most things can be reduced to simple, one-decision steps, and when you do that things will generally convert better for you.

Elephant in the room is payments. Was sad to not see this mentioned. Sure, building a web app might have a better chance of onboarding folks. But now try to get them to enter their credit card number. On mobile, they tap the "Buy" button.

That's fair. I do have something data about that, however. We tried making our app paid and found that we were now trading 100 free downloads for 2 at $0.99. So now we have 5% of 2 people instead of 100. Regular people are really price sensitive and the problem is they can find something that does something similar to what your app does for free. Like I said, they will not value their own privacy when making that decision, hence they will always choose the free version on mobile. I think in that future that will change, at least that's what we're betting on.

An excellent, excellent point.

From my reading, you are going from App-first to Browser-first, and less so from mobile-first to web-first.

I think it is critical make the subtle distinctions. We are in the business of serving doctors. For most doctors, having a native app makes very little sense. And yet, having a responsive mobile-friendly website is becoming more important than the desktop websites they presently have.

I agree that it is critical, but don't think it is actually a subtle distinction. It is a huge, glaringly obvious distinction. Which makes it all the more unfortunate that so many people reinforce incorrect terminology. I find it hard to believe people seriously don't realize their phones can access web pages, and that web pages can be designed correctly so they are accessible on all platforms.

Heh, he describes this long grueling conversion funnel, with bizarre steps like creating a group with no one in it. Their UI shouldn't even allow that. Reminds me of the founder of Path speaking at a recent conference. He said they were coming from the web and they had all these pre-conceived notions about how to interact with their users. When a user posts a photo on the web there's a whole process of getting it from their desktop or wherever, tagging where it was, tagging their friends, etc.. He said with Path they had to throw all that out because mobile users don't want that. They want a single push button remote control for their lives on their phones, and you don't put complex steps in that.

Another kind of strange comment was about how ad/virtual goods companies make most of their money from poorer, less informed clients. That's simply not true. Often half or more of the profit is from the whales in freemium games and a lot of stuff is put into the game just to support these whales. In one game with virtual goods that I write, I make a heck of a lot more money from the diehard fans that buy multiple $5 coin packs compared to the free users who slog it out earning those coins through play, or the cheap tourist who just uses the starting coins or maybe buys once and never again to try out playing with full bonuses for a bit.

Link to the path talk?

Read your entire article. If you weren't so parsimonious with words, I suspect I would be reading well into the night. :-)

More seriously: It is a given that app stores primarily generate free content for the platform-owner and enhances the value of supported platforms. (Additional outlays for dev+test hardware+licensing+developer-hours add insult to injury) Surprisingly not many developers seem to care or account for. With the amount of free noise in there, it is not surprising that you guys experienced what you did. Thanks for confirming what we had deduced indirectly.

Our startup launched deliberately as a web-app even though it would've been easier to get more (initial) eyeballs with a mobile-first strategy. For reasons you develop (and more), mobile apps aren't part of our current road-map (although we do address form-factor related nuances). We did consider expending resources to go mobile a few weeks ago but it made more sense to focus on creating new "content".

Last, the level of control gained (as you allude to) by owning A-Z of the product+distribution is truly empowering. That said, there is an undeniable development and eyeball cost associated with it. We should acknowledge this.

I think you need mobile. Leaving it out is a mistake because you need to be where your customers are (to me - of course it depends on your app). I just think that mobile is the wrong place to start right now.

Sorry for the word overdose. I have a degree in philosophy and computer science; just the right combination to produce overly lengthy and verbose blog posts.

1. Hope you didn't keep your response short merely on account of my comment...i was only jesting.

2. A huge fraction (imo) of the apps available in the app-stores did not need to go native mobile.

Mobile app-stores have mountains of apps that really have nothing mobile specific about them. This astounds me. Yet developers spent serious time/effort building these native apps. Why? I suspect it is the web-noise that continually dangles stats about gazzillion smartphone app downloads that triggers a greed hormone in philosophers and scientists alike, which begins to make them see mobile everywhere.

I am not suggesting that there's no need to address the form-factor issue. Only that "native mobile" of the app-store variety is frequently not a good strategy. Certainly not, as you point out, as a first-option.

Aside: You state "you need to be where your customers are". Yes, but if you define customers as those that play a role in generating revenue (either by paying directly for services rendered, or by permitting you to arbitrage someway), is mobile the place to be? This is obviously only a rhetorical question...you address this in your original article. I am just a bit outraged that the ios store has a million deluded worker bees building apps for it when they won't see a penny. Instead, they could've leveraged their work by owning the entire stack.

It's easy to make money on either the App Store or google play - work freelance for other people.

So Nigerians get the jab again.

Dear HN, how do we get rid of this reputation?

Hello Kind Sir,

I have great proposition to help you rid this undeserved reputation. To begin, please mail me soonest your telephone contact and address as well as account number so I can deposit funds to assist in our work. Once this has been obtained by me we will work together greatest outcome deserved.

Sincerest Regards, John Williamson, III

That was kind of insensitive considering the positive question hat he asked. Nigeria is a huge vibrant economy.

I guess I was just trying to illustrate in a creative way why people keep bagging on Nigeria. Once the spam stops people will forget all about it. And it's not just email spam, eBay and Craigslist are overrun too.

Sure it's a bit unfair as all countries have spammers. In fact, US and UK outrank Nigeria for scams, but those are two huge places and have a lot of other notable things, and Nigeria is #3.

So I think if Nigeria boosts its economy and provides alternatives for people to make money then the spam dies down and people forget all about it.

Well I'm just pointing it out : he's a tech entrepreneur and I'm sure he's really sick and sad to have that brought up all the time.

As punishment I ask you to go check out the tech hubs and the start up scene there http://cchubnigeria.com/about-cchub/ http://techloy.com/2011/06/27/5-reasons-nigeria-needs-tech-i...


Nigeria has 175m people, GDP grew 6.5 pct since last year. They have huge amounts of oil money and they are the financial hub of west Africa. Large middle class, skyscrapers, online shopping, massive film industry.

One day soon you will probably be trying to optimize your on boarding of Nigerian middle class users.

and 419 scams are an old tradition dating back to the 60s when they would write letters. So don't knock traditional culture ;)

Ha! Nice punishment. Thanks man.

I actually joke about it sometimes but it is really messing things up for legitimate people.

The fact we do not have other renowned stories makes the spam one stand out so much.

Rocket Internet did an online clothing store in Nigeria. Not sure how well its done. They are famous for cloning sites into different foreign markets. German company. Not the most honorable business model just to clone, but it pays the bills.

Highlight the varied and vibrant Nigerian economy. Publicize 419 prosecution efforts with a great well designed web campaign.

But unfortunately the 419 jokes are too often funny, so it's going to be hard to ever get rid of them.

I hope to visit the Lagos tech scene / hub some day soon.


Either Nigeria needs to be globally known for something really amazing (like a center of innovation, which will take a very long time to be recognized), or individual Nigerians need to walk away from the "Nigerian brand" and all that it stands for in the eyes of the global community, which unfortunately, is not much more than 419 scams.

Sometimes when a brand goes sour, or it no longer fits with long-term goals, it's better to walk away from the old and create a new one.

> you can test easily, cheaply, and fast enough to make a difference. > You can fix a critical bug that crashes your app on load 15 minutes after discovery (See Circa). > You can show 10 different landing pages and decide in real-time which one is working the best > for a particular user. You can also close a viral loop > . . .

I've always been skeptical of mobile-first, but so many people love their phones, I'd say if you know how to solve any of these problems (or the others listed in that paragraph), then that's a business right there. What an amazing list of pain points for anyone who wants to sell shovels!

I’m certainly not clicking that many Google ads per year and neither are you.

Are you sure? I used to think that, but I wrote a chrome plugin to test the hypothesis and found I clicked at least 1 or 2 ads most days.. I do 30+ searches per day. As a heavy internet user, you may click less on a particular ad, but you see a lot more because you search a lot more.

"I usually know where content stops and ads begin, partly because I’ve had access to the Internet longer than others. I think ads target the same audience that Nigerian scammers do"

It is hard to take someone seriously when they begin by claiming that ads are not useful to any party, and then gives a treatise on ads. But what's most offensive the fundamental assumption that only stupid people click on ads. (see references to "lower income" and "lower education"). Not to mention switching between the personal and the accusative ("We retain 5%" / "you're really paying $20").

My summary, I continue to be amazed at how important advertising is to the Internet economy, but that said, I understand it very well. Smart people click on ads. Smart people buy ads. Smart people build ads products and smart people measure and validate their effectiveness. If you go in to the ads or an ads-funded business with these assumptions, you will not attract good ads, and will get the incredible low RPMs you believe exist.

I am thinking along the same lines.

If OP reads this I would like to challenge him to count how many ad supported sites or apps compared to paid he uses and from the free ones how many would he stil use if they all become paid.

Also this comes from someone that has not run an ad based business. I wonder what happens if paid app model doesn't work out and they pivot again. Suddenly free becomes the greatest model ???

Things are never white and black when it comes to internet business, every business model has it's own place. Let's not forget that the no.1 web company uses ads for monetization. Would people pay to have google ads free? Do you think they feel taken advantage off because of the ads?

So, why not build mobile SAAS apps and charge for them? It seems like companies like Evernote, Remember The Milk, etc. that charge a premium for mobile as a value add are making money there.

My guess is the next big wave of mobile-first startups will be charging monthly subscriptions through IAP to small, but loyal users. Like 37Signals, but mobile.

It would make sense, but for now Apple is restricting monthly subscriptions through IAP to only Newsstand magazines/publications

The real pivot here is pivoting from free to paid. The medium isn't an inherent shift in the business model.

Agreed - a natural shift (somewhat mirroring what App.net is doing vis-a-vis Twitter). Curious to see how it turns out.

Separately it looks like this argument carries more water when talking about digital-only/communication products? Because when considering apps/services such as Uber, HotelTonight and Cherry (I hate picking exceptions) that provide an offline product or service, mobile first seems to have worked just fine. Perhaps the true underlying argument is that mobile first works best when the LTV of your customer is massive.

Perhaps. Every startup, mobile or web, start with a non-existent customer base, so I don't think one way or the other spells "more successful" necessarily.

That said, mobile-first companies are certainly unchartered waters compared to their traditional web brethren. I think they're a bit ahead of their time.

One distressing thing I can see happening: traditional web going the way of the newspaper industry. We're already seeing it with Google. They took a hit in this past quarter's earning's report because their traffic skyrocketed via mobile, but advertisers wouldn't pay the same CPM they were paying on the traditional web.

Sound familiar? Newspapers hit the same dilemma. They went from getting easy $30k tickets for single-run full-page ads to $3k/month online ads (to $300/month mobile-only ads).

Of course, this is a pure advertising model, which is separate from the OP. So... I'll stop here :)

It all comes down to what you are building and how your target audience will interact with it. Being mobile first is definitely hard when you are trying to penetrate into a sea of tough competition in the market especially if there is better use case to be web-first. For you, I feel you gained enough traction and very positive response by being mobile-first. I am sure that helped you get 1M or so users quickly on board and now that you want to go from 1M to 10M, you obviously need to think of entering new markets and platforms while maintaining your lead in mobile. This becomes hard and frustrating. We are a mobile-first startup trying to bring business cards onto mobile. It wouldnt make sense for us to be web-first, hence we are mobile-first.

I think you have a point with the registering user retention problem. Nothing is more painful than typing a password on a phone (especially if you use strong enough passwords).

But this very first step (I'm talking about choosing a nickname and password, not the social connection step) is clearly something derivated from what we do in browser apps. We could do that better on mobile apps.

For example, launching an app for the first time could generate some random login and password, and let the user instantly play with the app. When we consider he had enough time to get the picture, we notify him he must choose a password to be able to login from other devices, and he should probably choose a nickname to customize his experience.

That would be way more mobile friendly to me.

I think many of the people who are deploying through the app store would be better off deploying through the browser. Mobile web apps are surprisingly capable. I've ported one of my desktop web apps to mobile using appcache for offline support, and it works great. Especially on iOS where a homescreen web app launches chromeless there's almost no difference to a phonegap app except for the hardware / api access, which i didn't need because as pointed out i was porting a web app.

Yeah, I kept reading down to see where he would clarify that by 'web-first' he meant 'mobile-web-first' but that doesn't seem to be the case.

The tools for building mobile web apps aren't as good as they should (and hopefully will soon) be but it's certainly possible and gives you the flexibility to rapidly iterate, get the user up-and-running as fast as possible, etc.

I do not understand how the author's words relate to the topic.

Mobile first design is designing core functionality og websites for small screen, big target devices first and using techniques like progressive enhancement / responsive design to scale up for desktop experiences.

Mobile first already means web. Are they going from mobile first to desktop first?

Or do they mean mobile apps to web apps?

These terms have established meanings. By sticking to them we help simplify our discussions.

I agree, not with every point, but with your conclusion.

My company (PicDigest) is betting on a paid web-first model and planning a mobile app as a companion, even in the crowded space of photo sharing where incumbents supposedly have the market sewn up. Yes, it's a long-shot play, but the wager is based on the assumption that many people will pay a little money for what I'm offering.

The statement "I’ve had access to the Internet longer than others" sure reads funny coming from someone born in 1988

Even if that wasn't true, which it is (there are still billions of people without internet access), I've spent the majority of my time on the internet and I think I can tell the difference between ads and content.

Making this required reading for startups before they talk to me again about mobile and other non-revenue ideas.

Why not build an HTML5 site? You'll be device- and platform agnostic, and you can still iterate quickly.

Didn't FB try that?

Yes, but don't jump to conclusions.


As others have been pointing out feverishly on Twitter: the problem wasn't them betting too much on HTML5. Their problem was developing piece of shit apps that happened to use HTML5. They tasked amateurs who didn't know what they were doing into building a hybrid native app container which in turn embedded HTML5 content. Plenty of other developers (Instagram and LinkedIn come to mind) have figured out how to do that right, and in a way where it is seamless to the end user and for all intents and purposes feels exactly the same as a native app. I'm not saying that it's an easy problem. You have to find the right balance between which components should be native or not. It's clear from the other problems that Facebook's been able to solve that they know how to hire top-notch developers. They just failed to do so for their mobile efforts, which just reinforces the stereotype that they don't "get" mobile.


Yes, and it is more used than their mobile app. Were you referencing their use of html in their app? That is an entirely different issue, but yes they also did that.

The term "pivot" has clearly "jumped the shark".

I've also viewed "mobile first" as thought/design methodology that doesn't mandate you must then "build" mobile first.

What's with the word "pivoting", anyway?

I think I've seen it used more often on HN in the last month than in many years before that...

It's a matter of what your product is, really. If you have an innovative product that solves a specific problem or need , then it can work. It doesnt work for all products. Mobile development wether it is native or a web app , is hard ,and yes you have to go through all these processes on some plateform. Is there some money to be maid ? i definetly think so , but it is not really an el dorado with easy money. It is easier to make money on the "desktop" web than on mobile right now. Things may change in the future.

Applications are open for YC Winter 2024

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact