Convopage
See the entire conversation
Why would someone add this banner to their website?! Embrace the web it is good. Tumblr web is great.
153 replies and sub-replies as of Apr 10 2017
reddit does the same :/
There's a funnel problem and Android-blindness. Both loom large in these conversations.
Many product managers "know" that native apps have higher engagement. When they didn't with them, lo and behold! It's true.
What they miss is that they aren't running a reasonable comparison unless they've also built a PWA.
ugg...auto"correct" killing me. Anyhow, when they run native app experiment w/o PWA baseline, they select for *already* most engaged users
The set of people who are willing to sit through a native app install, updates, etc. already find value from the service and want more of it
Mind if i steal that quote for later?
Sure!
I see a lot of confirmation bias: the set of people who will jump through the high-friction install is low, but they are engaged
This plays into the Android-blindness: in the US and many other developed nations, ~100% of those PM and business folks carry iOS devices
...and iOS sets your sights low for the web. This causes terrible cognitive dissonance for folks trying to grow their businesses.
For many, it has been a tough shift to think "mobile first" and in developed markets, the rich people *also* have iOS devices.
If you're looking at revenue growth in your existing (developed) market, you might dismiss the web offhand or relegate to desktop.
Lots of companies we talk to have gone through this cycle. They built for the web, tried to do responsive w/o rewrite or scale-down
...predictably found out that people didn't want these bad experiences. Tried native apps; got some good data from most engaged users...
...then hit a wall. Native app distribution flatlines very quickly unless you keep pumping marketing and/or UI degrading prompts into it.
But for many teams this isn't a (short term) problem. They are in different parts of the org: native apps either outsourced or in other team
...and frequently funded out of speculative budgets (marketing, R&D, etc.). This deprives the organisation of rapid learning.
So they learn on long timescales (a.k.a. the expensive and hard way) that these sorts of prompts can "work" but in a highly contingent way.
Luckily we now have PWA tech in browsers and companies/brands can run *real* A/B tests. What does that look like?
First, you want to track the total funnel from a prompt. What's the *overall* CTR for a web users to convert? Are launches the same?
To get this data you need to invest in PWA-ifying your website. Users adding non-app-like experiences to their homescreens is apples/oranges
Confoundingly, Web Push is unbundled from app install. When we tell folks this it takes them a little while to get it, then "ooooooohhhhhh!"
This is *great* for users, but makes the A/B test much harder to do rigorously.
Imagine you're the Tumblr or Reddit PM: what's your *goal*? Is it to tell a great story to your boss/board about the # of installs?
(remember that installs aren't a business metric; they may have zero correlation with profit.)
In SF, "installs" too often viewed as a strong proxy for growth. Savvy folks know better, but definitionally there aren't many savvy peeps
You can often get promoted for a long time by driving up installs even if all you're doing is strip-mining user value & slowing development
Small companies too often do this organisational learning in a darwinian way. Big companies outside of tech can frequently apply more rigor.
...but they tend not to be as familiar with all of the options. And there's no marketing department for the open, interoperable web.
Very few businesses have run real A/B tests and fewer will share the comparative native app/PWA engagement/cost data.
(for obvious reasons; it's either competitive advantage, potentially embarrassing for the folks internally who pushed native apps, or both)
We have an inexact proxy in the cost to acquire a user installation (the top of the "install" funnel of engagement). Those #'s look *great*
Housing.com's cost to acquire a PWA user is 50x lower. Selio's was (only) 10x lower.
Look Up a place to Buy/Rent in India | Housing
Search Property in India's most trusted Real Estate Portal. Browse New projects, flats, ready to move apartments for sale #Housing.com.
housing.com
It's going to take a while for the big companies that can both invest in PWAs and do rigorous studies to feel comfortable sharing, I fear
...and of course this could just be wrong! It could absolutely be the case that there's some property of native apps we've totally missed.
And there's also the very real possibility that the folks making PWAs will just drown the experience in script so it'll never feel good.
...which in a different way deprives us of real A/B examples. Web developers are still struggling to take mobile seriously.
So that's what we're up against. People putting terrible prompts on sites that could just be PWAs are doing what they think reasonable.
...and for iOS (which takes up a disproportionate amount of mind-space), it *might* be reasonable.
It's complicated!
The prompts and dark UI patterns (*ahem*, Messenger tab in FB on mobile) are bad for users & degrade experience of the web.
...but we won't end them until we convince businesses that they're *also* not good for the bottom line.
I think you should put all these ideas in a terrific medium post we could find later 👍
Yes, please!
Yep, +1
I second this motion. Great insights!
curious about these numbers. is there somewhere I can learn more about this?
This entire thread is 💯 btw. Really insightful, would read again
To be fair, historically native apps have had positive attributes webapps haven't. Need to build good PWA to see equivalence.
can push permission be automatically granted on install especially with WebAPK?
We're looking at that. I'm not convinced it's a better user experience.
Probably. Not saying it's 100% good. On the other hand, business will have more reasons to make PWA because that will be 1 (a2hs) prompt
Instead of 2: a2hs + web push
That assumes business need more reasons. But I take the point.
That's wrong phrasing probably. "Business will be more convinced" maybe. All reasons are there, tough to explain it some times.
Also many managers who managed native app previously start managing PWAs are demand push perm request immediately on PWA load
also I had a request about "on-boarding process" for PWA.. so..
Yep. It's a change. Do you talk to them about push sub *before* install? Not possible on native, big advantage.
I showed them @owencm's UX guide for web push and they agreed to it. Would have hard times explaining myself otherwise
It's easier to convince clients when something was written "by the Google who invented PWA"
Fascinating and a little sad. Might have been my plan, but wouldn't have happened w/o Moz, Samsung, and many, many others!!
I know, I know. I have an article about that :-) (that one without "fact" 😄)
Progressive Web Apps isn’t a Google-only thing – Arthur Stolyar – Medium
Progressive Web Apps isn’t a technology, but a terminology which describes rich website in a way of quality and ability to progress as …
medium.com
You're missing the @samsunginternet logo I believe.
That was long time ago. Also image was googled and I believe I used that image to show major browsers. e.g. UCBrowser also not included
Woah thanks for the amazing thread!! Sorry didn't reply was stepping onto a plane when I tweeted. Landed to over 100 notifications!!!
Do we have examples of companies doing PWAs to compete to their native apps?Most are "Lite" versions. They are afraid of taking the big step
Most of the "Lite" branding belies (near) feature equivalence: Twitter, Lyft, Flipkart, etc.
Seriously, try both native and their PWAs.
And "lite" refers to the user value of not needing to wade through an app install. Also a nice way not to raise hackles of native teams.
Remember that PWA teams inside orgs are often _insurgents_ against an all-in-on-native strategy.
Its a shame I can't even get to the Facebook website any more on mobile, it auto navigates to the app
If you uninstall the app, should be accessible?
The problem is I used the web for one account and app for another so that's not an option 😭
That's why we need to find a way to give them real metrics to convince PMs to go Web. Love your Wego example Is their any written on this?
I don't think users understand that idea of "Lite". That's a value only for users with low level devices or countries with bandwidth issues.
Twitter is promoting the PWA as Opera promoting Opera Mini, not the full Opera browser. And Twitter is not the only one.
It's turned on for all of mobile.twitter.com. Flipkart "Lite" is all of Flipkart.com on mobile, etc. Watch what they *do*.
Online Shopping Site for Mobiles,Fashion,Books,Electronics,Home Appliances & More @ Flipkart
Shop for electronics, apparels & more using our Flipkart app Free shipping & COD.
flipkart.com
Flipkart is great example. But they still promote the installation of their native app within the PWA. It's like two steps forward, one back
Where? I don't see it.
I took some screenshots when I saw those app banners. I'll search for them in my android devices and share. Probably not today 😎
You're on iOS...It's reasonable for brands to try to get installs there; Apple blocks meaningful PWA support 🤷
I saw the banners on Android. In fact it happened live on a conference. Not sure if it's still there or not. But it was an unclear message
Good businesses experiment. You might try asking @nagarajuepuri, tho.
We are in the process of completely removing traces of "App Install" page and killing it. If you saw one previously, it's due to tech glitch
We strive to provide best possible experience wherever user lands(including all features) "Lite" or "No Lite" is just a name of product
I wish every org follows this and takes away of user perception of "Lite". We r in an era where providing best experience across is possible
I believe this is just a matter of time before everyone realises this fact and start delivering near same experience across all channels
we're still missing some very basic, *truly* cross-platform, App-like features, but I agree we're definitively moving to the right direction
I'd agree with you if once there using safari I'd be asked to install their app too. That's not the case though, they probably have old card
Double 🤷
I still use FT as a good example, they took the risk went PWA only on iOS back in the day to incredible success.
They supported a wide range of platforms with a small team back when interop was much worse even flexbox was an enhancement.
On top of that they charge a subscription to use it! Showing that people were willing to pay for a web experience, against the common wisdom
All of this during a time when there was no service worker, no install banners, no ambient badging, no push notifications.
I know all about it. I wrote about it before the guardian article. What's the point? It wasn't enough. They were pretty much alone for years
I'm not saying we have a tech problem. I'm saying we are over reacting on how companies are trusting the Web Platform and PWAs.
If we don't accept the problem we won't find a solution to it, if there is a need for a solution. Hiding the real situation doesn't help.
I literally have no idea what you're talking about. You make progress one improvement at a time; that world isn't perfect != not improving.
OK. I think it's enough for a Sunday Twitter discussion. But the PWA approach needs more action and less talking if it wants to succeed.
We need more Google (replace with other company) replacing a native app with a PWA. If it doesn't happen, companies don't trust enough.
Many people confuse having succeeded with the grungy work of making progress a reality.
I was on that team. It was very talented team with some great testers. I'm not saying it's easy. No success is easy but it is inspirational.
Google pre-installs its apps, hence the proposition of low friction makes less sense for us. It's unfortunate, I'd love to see us lead 😕
Oh I am just saying that if it could be done then. It should be easier to make a successful PWA today.
FWIW, I asked them about their Twitter card config last year. It's from before their PWA work, but not clear why it hasn't changed.
This is because of Twitter App Card ( dev.twitter.com/cards/types/app ) included from the desktop site. Legacy behavior.
If you saw that it was during an experiment we had to run, the results of which convinced enough folks at Flipkart never to do it again ;)
Good to hear that. It was a weird message.
It's sad thing that showing native app banner is industry enough and every company and manager has to be convinced not to do it.
Can't argue without data.
With me? Why argue with me. With them? Yes, sad thing.
Industry enough -> Industry standard
Yea, twitter "lite" has no user visible branding AFAICT. Dev blog post won't be seen by most users.
I'm not saying is bad. But we need big players taking the big step before saying PWAs are worthwhile. If not, no one believe us 😒
I think nobody in this thread is saying the contrary or not believing on that.
Since PWAs are web, they are also less of a security risk than native apps, right?
To the extent that web apps are less of a security risk than native apps, yes.
Speaking as a random user with a mild understanding of all things PWA; Twitter PWA is awesome! GG Well done.
Can you share more adoption info? I am surprised so many more people choose twitter.com on mobile over an app install.
However I am in the USA so it's obviously different in other parts of the world
Been using it for a few weeks now and love it. Add list management and I'm all in.
"More people experience Twitter Lite than any other Twitter client" wow, I would not have guessed that at all
If they were starting today, I bet they could just write PWA and do not write native version. Then just package with Cordova for the store
True. 'Lite' has little do with feature sets.
I didn’t say it has to do with that, but just seeing the definition most people use
The 'low-fat' definition seems better :-) Hopefully metrics should start informing opinions soon.
But do the users understand that?
It's less of a mouthful than PWA, and if twitter doing a 'lite' ver gets people building PWAs then I think that is okay.
I know but they are not convinced, they told theirs users this is not the full app. With that happening, difficult to make a fair comparison
* Lite -- better scrolling
* Native -- sidebar drawer works better
All the difference :-)
You should check out Flipkart lite side drawer. We spent a lot of time getting that right!
I did, it's great! But you may want to checkout Polymer's drawer: at wego.com or polymer-project.org/2.0/start/
Apples to apples would be native app with 'Added to homescreen' metrics, given a well executed PWA.
having a hard time understanding the last part here. Typo I can’t figure out maybe?
yeah, sorry, auto"correct".
s/didn't/experiment/
thanks!
This could also be an ad revenue thing. More money in native ads that can't be blocked and aren't limited by what a browser will allow.
bleh 😥😝
Also I think it you can only browse for a bit before it asks you to use the app or else 😳
How to build webapp in 2017
- Make a really really great webapp (PWA if possible)❤️
- Put a big "Install the native app" banner
- 🖕🏻
Good thread, pasted all up in our Slack, thanks much.