Hacker News Reader

Comment by ricardobeat | Hacker News Reader

This is what https://www.hopper.com/ does. Didn't quite work out when I tried, prices only kept going up!

Hi there, Hopper engineer here. The blog post is pretty accurate from my experience (lowest prices ~3 months out) but if you apply any of this logic to the past year then you're going to have a bad time. The pandemic caused the entire industry to lower their average flight prices and now that things are starting to return to normal, so are prices. Because of that, predicting price trends based on historic data is a bit tougher.

If you use iOS (not sure if we have it on Android just yet) then I'd recommend looking for something called "auto-buy" when you search for a flight. It's basically a limit order for flights and it'll allow you to set a price that you would be comfortable buying at. If the fare dips to that price then it'll book for you. We've seen some insane deals get booked through this product. Just looking at some of them from yesterday there was a round-trip flight from NFO (Norfolk, Virgina) to MCO (Orlando, Florida) that we originally price quoted at $284 for the user and it was booked at $166.

Why do you have some features on ios but not android? Why isn't this on your website.

Maybe I am old (possible), weird (probable), and out touch (definitely). But I couldn't imagine the case where I am planning on doing research for travel and I would go to an app over a website. For looking at data, typing things in, comparing prices, etc. A phone app really is the worst form factor for that.

I am not sure how big your company is, or how its done. But your product owners are dropping the ball in my opinion. I backed out of the website as soon as I found I couldn't do anything from the website.

I guess I'm also old, weird, and out of touch :) I cannot begin to describe the annoyance I experience with filling in forms on a mobile device, especially if I'm hunting for a good deal `to -> from` and need to keep going back and slightly modifying my search params.

Phones just aren't good at this use-case.

Being (very) cynical, the only reason they're providing a mobile app is because of all the data they harvest (https://www.hopper.com/legal/privacy-policy/).

Thankfully there are millions of people that prefer to use their phone for searching and booking travel and we've been able to build a business around it.

I feel your pain around filling in forms and managing context on mobile devices which is why we've tried to make the process simpler and deliver to customers what's actually important. This article was written by our head of design back in 2018 about why we chose the search parameters we chose for flight search: https://medium.com/life-at-hopper/users-dont-want-filters-th...

He's now leading design at a start-up called Fast trying to apply similar principles to paying for things online.

And who knows how many more untapped because they don't want to use your app?

Every time I've (tried to) use the RyanAir app to book (or browse prices for) a flight, for example, I've given up and used a 'real' computer to go to the website. That's not (just) because the app's bad.

Having app-only features and limited functionality website is just infuriating. Things go wrong with phones, websites are sort of a 'lowest common denominator'; you're not locked out because your phone broke and you had to borrow someone's old one but it's $OS and you're used to $OTHEROS, etc.

> Thankfully there are millions of people that prefer to use their phone for searching and booking travel and we've been able to build a business around it.

How do you know? It's not like you gave them the option. Anyway, like you said, I'm not your demographic and I realise you are helping a lot of people travel more cheaply so I do wish Hopper all the best :)

Yeah, same. Guess the old saying "if something's "free" then you're the product" hits home here too. Forcing the mobile app is a deal breaker.

While that saying applies for things like social media and such, it isn't the case for Hopper because you still have to pay for travel. The app is free, but you still have to pay for any flights or hotels that you book. Usually when an app or online service is free, then you'll see ads serving you content associated with your search behavior because that's the only way that service could monetize itself. Hopper's founder has been pretty adamant that we'll never serve advertisements in the app and we wouldn't be selling user data to 3rd parties and at least in my 4 years at the company that promise has been kept.

I would hand in my resignation if we started serving ads in the app.

This. I also lost interest as soon as I saw the website just re-directed me to an app.

Not sure how much I can say, but it mostly comes down to engineering capacity. We're still a relatively small engineering team (100-200 engineers) and for a long time were more like 30-50 engineers. There's a lot of moving parts in travel and we chose to focus on iOS mainly since it's the has the biggest market share in the US for mobile.

Hundreds of engineers and can’t build a website. Something is wrong here...

It's likely you're underestimating the complexity required to run Hopper and support all existing features. Only a small subset of those engineers will do mobile work. And they likely have many more priorities than building Auto-buy, the benefits and viability of which isn't obvious. So while the rest of the Hopper Eng team churns on with other long- and short-term changes, one or two engineers may have been able to lobby for a POC on iOS. Or maybe it was just a quick week long hackathon project that's still being monitored for ROI.

Something is potentially wrong. Saying "we only have 100-200 engineers, we can't get around to making a website" is a terrible excuse that points to mis-management. It doesn't take more than a couple developers to put together a basic "shopping for flights" web interface. If they have a backend that serves iOS, it should have a REST API, so slapping a web interface on that shouldn't be a big deal.

But it's also possible that it's like someone else mentioned: "if the product is free, then you're the product". Maybe they just don't care to get more customers/solve their problems, because they are harvesting enough data about iOS visitors to make a profit. If they don't want any more customers, then nothing is wrong.

I am young, and decline to comment on the others, but fully agree. It irritates me to try to do such things on my phone; I often give up and reach for a 'real' computer & keyboard to start again and wish I hadn't bothered with the phone to begin with.

If I didn't need a WhatsApp app on a phone in order to use WhatsApp on a computer, I probably wouldn't have a phone, (maybe an emergency use classic Nokia brick) or at least use it extremely infrequently.

Why do you not allow for functionality on your website, but force users to use an app?

They actually tease you by pretending to show results on their website [0] but when you click on any of the flights it just takes you to their homepage.

[0] https://www.hopper.com/deals/best/from/city-toronto-YTO/to/c...

I'm impressed you even found that. AFAIK there's no link to it anywhere from hopper.com. It was a little SEO project that one of the engineers that's been here a while worked on using our internal prediction APIs. Something like it existed in the app at one point, but the engagement we saw with it was pretty miniscule.

Appreciate the kind response. Those pages are linked right on the homepage, I clicked on the Paris[0] one.

[0] https://i.imgur.com/7ey6spH.png

Oh wow, goes to show you what I know... I assume when we were working with the company we contracted out to build that fancy website they asked if we had anything to link to for flight deals that we were just like "this engineer had a little pet project you could link to". The funny thing is a new hire just asked why we didn't have anything in the app for "best deals in general from my city" and I responded by linking to that project.

It used to exist as its own separate app called "Get The Flight Out" but for some reason we removed it from the app store. We also had a feature devoted to something similar for "flexible watching" but it never got much engagement and it didn't warrant the infrastructure cost it required to keep up.

yeah... i find that a bit disgusting.

Assuming good faith, the business decided that mobile apps were more valuable to their customers so they've spent their (inherently limited) resources improving the mobile apps. Assuming bad faith it's because they want to get whatever data they can off of you while using the mobile app. But due to Apple more clearly exposing what permissions an app uses, that is not quite the concern it used to be.

I'd bet anything that the business didn't decide that a mobile app was more valuable to their _customers_. Maybe more valuable to them, but not to their customers.

It is very rare that an app will provide more value than a website. And most of the time when it does, it's just because the website is shit, and they spent more time on the app. But if they focused on the website, it would be way better than whatever experience the app is giving you.

Here's something to think about: would you download and install a desktop application for every random website you come across? No? Then why do it on your phone? People must just make better websites.

I'm assuming good faith here, and you should too until you find out otherwise. It's fine to be skeptical, but it's rarely helpful to be cynical.

Yea it's not like you would ever use your phone for anything that you couldn't just use a website for. Before the pandemic when I was at a bar and needed to get a ride home I would just take out my laptop and find the number for a taxi company to call to pick me up. This is the perfect solution to the problem and a mobile app could never be more useful.

Yeah, the parent comment is unnecessarily cynical. I've worked at multiple travel companies that have independently either built an app or considered it based on user feedback. I'm not the biggest fan of the idea but I've come to peace with the idea that other people interact with their devices differently than I do.

I wasn't saying never use a phone when you can use a website. I was saying why have a dedicated phone app when you can just use the browser app on your phone. Or are you saying you downloaded an app on your phone just to lookup taxi numbers?

I'm saying there are 2 multi-billion dollar companies that only had mobile apps for a long time that the vast majority of people use to book rides to places.

Because there was a vacuum in the mobile travel space when Hopper was trying to find its main value proposition and we tried to fill that vacuum. Millennials (and now Gen Z) tend to be pretty price sensitive and they also tend to be a higher share of the mobile user space so it made sense to apply our "you can watch for flights and get notified when the cheapest flight hits what we think is the lowest price" feature on a platform where sending notifications that users actual on is actually feasible. On web, the options are email or web notifications and neither of those are very actionable in a timely manner.

They did preface the post by saying they’re an engineer. Most likely that wasn’t their decision.

I fully expect that it wasn't their decision, but also presume that they'd have insight into why a decision like that would be made.

Also, it allows me to express my disdain at such a decision and if enough other people did too, said engineer has the ability to go up the flagpole and inform the decision makers that there are many people who are displeased at their decision.

I do have that ability, but I also have the sense to know that HN readers is not who we're building the product for and when you look at the travel industry as a whole the SEO/web advertising game is not one you really want to play. Two of the biggest spenders on Google AdWords are Expedia and Booking.com. They spend our annual marketing budget in every like 10 minutes or something like that. We literally can't compete, especially when you consider that Google is doing their own with Google flights and hotel search which favors themselves.

So the smart thing to do (at least in our opinion) was to not even play the search engine advertising game and focus on building mobile apps that users will come back to after using. We tend to see really high retention rates with users because they like the experience and we don't have to pay for the ad impression every time they feel like searching for travel destinations.

To make more money. Like most "dark" patterns.

Just a thought, perhaps it's something to prevent scraping.

This actually is one of the benefits of being a mobile only product! It makes it a lot harder to scrape the data that we produce (prediction and such). I don't think that's ever been the main reason though, just a side benefit.

Why don't you have a web interface?

Just want to say that Hopper is fantastic. Love the simple app design and I’ve gotten some great deals with it. I’m at least one HN reader who does a lot of travel planning on his phone.

The only time I remember the predictions failing badly was booking a flight for the 2017 full eclipse. Hopper kept advising that prices would fall, and no surprise, they never did. Obviously historic data was little help for a rare social event like an eclipse.

Thanks for your answer! This was maybe 2-3 years ago, way before the pandemic hit. I started watching a flight - probably AMS-SAO - a few months in advance and all I got were alerts for the fares going up every other week. Pulled the plug at some point with the usual [expensive] fare. Maybe I was just unlucky.

Norfolk's airport is ORF.

Some cities have their own (not really official AFAIK) codes that represent "any airport in this city", such as "NYC" meaning LGA, JFK, EWR, and maybe also HPN and/or SWF. I think these are mostly created by app/site developers but I could be wrong.

They're metropolitan area codes (https://en.wikivoyage.org/wiki/Metropolitan_area_airport_cod...). They're not consistently useful across sites because different sites connect to different backends (GDS) that may or may not accept them, but it's nice if you consistently use a site that does. Back when Hipmunk was alive QSF was useful to me because it used Sabre as the GDS and would translate it as {SFO, OAK, SJC}. Other sites would show you an airport in Algeria.

Edit: I should note that you should be careful when using these - at least once I accidentally booked a ticket that left from SFO and returned to SJC.

That's the term, it's been a while since I had to work with this stuff. I don't remember if I ever was able to use the metro codes to look up data from a GDS, but I do remember that I once ended up with a dataset where I had to heuristically disambiguate those codes from proper IATA codes; that sucked.

That would require Norfolk to be big enough for two airports :/ The next closest is Williamsburg/Newport News and then Richmond.

You're right, but it was just a brain fart on my part. It was ORF and I don't think Norfolk has a city IATA code.

A shame that the flight wasn't between NFO and MCO; I would love to book a ticket between Orlando and Tonga for only $166.

Ah you're right. Looking at that slack message again it was ORF. Not sure why I wrote NFO.

I'm guessing you had a pretty strong prior that 'prices go down'?

It got me LAX-JFK for $361 during Xmas holiday... usually it's like $600.

Reply on news.ycombinator.com