Google warns that rate limits, overage fees are coming to Maps API | Hacker News Reader

Google warns that rate limits, overage fees are coming to Maps API

arstechnica.com

101 pointsshawndumas83 comments

I remember in the 90s when we lambasted Microsoft for cornering markets with free products (IE) so they could drive out the competition (Netscape) and eventually start charging for them when we had no other choice. Luckily that never really happened, Microsoft never started charging for IE.

While Google isn't exactly doing that here, it's coming awfully close.

Though that's how it may look, I don't think that was the strategy from the beginning.

I worked as the support engineer for the Google Maps API starting in 2007, in the early days of the API. We didn't think about monetization then. But then the GFC happened, and all over Google, there was a bigger emphasis on monetization -- basically every product had to eventually monetize. We experimented with different types of ads, but as it turns out, map-based advertising is difficult -- it's hard to know why a user is looking at a particular map, it's hard to get advertisers to geo-locate their ads accurately, it's hard to squish ads on a map without adding clutter, etc. While we continued to experiment with ads, we started talking about "ala carte" pricing, letting developers pay-per-view, per-geocode, etc, as a form of monetization.

That was about the time I left Google, when the monetization strategy was still unclear. It looks like they managed to implement ala carte from the technical perspective and decided to introduce it. Though it is a harsh reality check for developers who are already using the Maps API on popular free websites, I think its probably for the best. If I was given the choice between no Maps API at all, a Maps API with (not-so-relevant) ads, and a Maps API with incremental pricing, I'd personally opt for the latter. And who knows, maybe map-based ads will become an option in the future for those who prefer that option.

And if the pricing doesn't work for your site, there are multiple open-source mapping libraries that could use some love to bring them up to par with the Google Maps API (both from the developer & user usability perspective). I mentioned some in this talk: http://www.slideshare.net/wuzziwug/open-maps-or-close-enough... ...but there have been a few new entries in the space since then.

(Note: As I no longer work for them, nothing I say should be taken as official word from Google.)

Two interesting things in the open mapping space since the talk was published are TileMill:

http://mapbox.com/tilemill/

And Justin O'Bierne had some great blog posts on why Google Maps work so well from a usability perspective:

http://web.archive.org/web/20110130154705/http://www.41latit...

Sadly, he's taken down all of his posts.

I agree with that, it is quite similar to what they did with youtube and the unbearing amount of ad that one has to endure before watching a video now, after 5 years+ of making it available free of annoyance.

Uploaders of videos can enable/disable the inStream ads. They are not required by Google. http://www.google.com/support/youtube/bin/answer.py?answer=1...

I don't have to bear one second of Youtube ads. (Thank you AdBlock!)

The "Shark Week" ad that literally screamed at me at 11PM was the straw that broke my mental back. I immediately installed AdBlock.

Some video ads (on YouTube and elsewhere) can detect AdBlock and refuse to play the video.

Thankfully there is an alternative now (that wasn't nearly as good 5 years ago): OpenStreetMap.

Their main *.openstreetmap.org has Tile Usage Policies, but there are other companies that provide free serives, and you can also host it yourself, render your own map, and pay your own bills. The data & tool chain is entirely Open, so you are not at the mercy of anyone.

Unfortunately OpenStreetMap doesn't really cut it.

I'm using the GMaps API for geocoding addresses and in that regard it's the best one available, especially since it has a pretty high tolerance for spellings and non-standard address formats.

The quality of OpenStreetMap data really varies a lot. It is often better than Google in european cities.

When you can't innovate anymore, you increase your prices. After a few years, the company becomes a dinosaur loaded with cash... like Microsoft.

In what kind of bizarro business model are price increases and innovation interchangeable?

What do you mean "cornering the market"? You can always use Mapquest/Yahoo Maps/Bing Maps on your site.

This has an MBA-in-charge smell about it.

I'm not certain how many GMap implementations there are that exceed a 25k/day load, but my back-of-the-napkin guesstimation is <10% of all sites that use Google Maps.

But, those are exactly the sites that I want using my maps, instead of Yahoo or Bing or some other provider.

Unless there's something I'm missing, this seems penny-wise and pound-foolish.

There's actually not a huge benefit to Google in my site using Google Maps instead of Yahoo Maps. There is a branding benefit for sure, but there are no links to Google services, no ads, or anything like that.

I've been wondering for a long time when this would come- the processing (getting directions, geocoding, etc) and bandwidth (for map images) demands must be huge, with comparatively little payback.

It'll be interesting to see how many big operations pay up, and how many switch. To an extent switching might be futile- if Google can't afford to keep the maps offering free, do you really think Yahoo can?

> I've been wondering for a long time when this would come- the processing (getting directions, geocoding, etc) and bandwidth (for map images) demands must be huge, with comparatively little payback.

I'm more thinking about the costs of developing the software and gathering the data. And also the perceived _value_ of the data, independent from its costs. Ever tried to buy the kind of data Google uses on their backend, with terms of use that would support the kind of services Google offers? Good luck--both with terms and then with pricing. Other companies have been charging a fortune for this data.

> It'll be interesting to see how many big operations pay up, and how many switch. To an extent switching might be futile- if Google can't afford to keep the maps offering free, do you really think Yahoo can?

Regardless of whether they can, with Google pulling back, don't you think Yahoo et al. will want to raise their prices?

Don't forget the cost of the actual data. Remote sensing satellites are not cheap. Google Maps is heavily dependant on a few companies who provide the bulk of the data.

They have driven cars around nearly all the roads in several countries taking photos of everything... They could use that to make better maps.

One advantage of Google releasing Google Maps for free was that now everyone & companies got used to having web maps, and hence places started to put their locations on the map (so it'll show up in google maps). There is now a lot more geographical data out there on the web.

" - the "JavaScript API or the static maps API will be capped at 25,000 map loads per day.

- They will be expected to pay a fee of $4 per thousand loads in excess of the limit.

- The cap is set to 2,500 loads for styled maps that customize the presentation.

- After 25,000 loads, the overage fee for styled maps will bump up to $8 per thousand. "

That is no joke, a $4 - $8 CPM on map loads means your app has to be doing substantially better than that, which is NOT an easy enterprise.

Not exactly. If you use 30,000 a day you will pay $20, or only $0.66 per one thousand.

They are giving you over 9 million requests a year for free. Its not unreasonable to expect the profit from that to help subsidize the excess,

Sure, if you plan on staying small and never growing, this is fine. But you can't build your site for growth staring at the edge of an $4-8/CPM marginal cost (in addition to all your other overhead!).

$4-8/CPM is far more than most ad networks pay. Perhaps Google is deliberately trying to get rid of most major GMaps users. If they follow through on this, they will be successful.

Sounds like they are trying to push their high-bandwidh users to enterprise pricing.

The $10k/yr Enterprise licence works out to $27/day which breaks even with the incremental pricing at 58k visits/day.

Isn't $10k/yr just the starting price of the enterprise license, don't they have tiered pricing there too?

The price I was quoted was actually $17k/year for 250,000 impressions. We asked for clarification, and the salesperson said yes, you only get 250,000 impressions a year.

That's $68 cpm.

With some pushback they increased it to 1M impressions a year, but it's still a bit ridiculous.

There is tiered pricing. Per client basis... with large discounts for large volumes.

Yeah, the pricing is absolutely ridiculous.

At least there are alternatives:

1) Yahoo maps

2) Bing maps

3) OpenStreetMap

4) Nokia Maps

5) Cloudmade

There's also (I'm as surprised as you are) Mapquest. They have a whole host of open APIs now, built on OSM, and even have web services like directions/elevation: http://developer.mapquest.com/web/products/open

http://developer.mapquest.com/web/products/open/directions-s...

Ahh Mapquest. I'd forgotten all about them.

They are one of the poster children for early success and then resting on their laurels while the competition ate their lunch.

Except MapQuest is yucky.

I believe Yahoo! is out of the maps API business as of September. http://developer.yahoo.com/maps/

Seems to me like Google has only two dials: free or expensive. For example, in the case of Google Apps, the jump from free to $50 per account / year is rather steep imo. $4 fee for each 1K over seems rather expensive. Anybody know the price range for the subscription?

I agree with you that Google's Apps pricing is steep, but I don't think they're marketing it towards the average guy who just wants a couple of email accounts for his personal site. I think they're marketing it towards businesses and schools who see $50/acct/yr as pennies compared to hosting their own email server, hiring a sysadmin (or telling their current admin to handle it), and then supporting that server all day, every day.

I'm not a small business owner, though, so I don't know how much one would generally pay for email, or how much $50/person/yr is compared to the rest of the business.

$50/acct/year comes to $4.16/acct/month. If your staff's use of email isn't in someway contributing to making more then $4.16/month you should probably just fire them, or not give them an email address in the first place.

I'm talking about places where maybe 10 people really need email and 100 other maybe once a week if that. You can definitely justify the cost for those 10 people, but not for the other 100.

If you could mix and match free and non-free accounts this would work great, but you can't. Again, I'm talking about companies with razor thin margins where the technology side is not super important. What do they do now? They mainly use free gmail/live/yahoo accounts.

You can in fact mix and match - you can use Google Apps on top of a corporate system such as Exchange for only specific users. It's unlikely to save you any money since you'll then need someone to run it, but it's possible.

As things stand at the moment the company I'm working for is in exactly that boat of some staff depending on email, and others only needing to retrieve a single email on each day they're working. We're working around it right now by just getting them to give us their personal email address.

That probably won't scale long term, but for now it does the job, and if we get to the point of needing more then I may just throw together a quick and dirty mail server for the people who don't need the convenience of Google Apps syncing between devices for them.

50 users @ 4.16$ / month = 208$ / month subtracting 50$ / month in HW and bandwidth you left your admins with about 3 hours of work @50$ an hour including overhead. Sorry, most admins spend more than 3 hours a month dealing with internal email at small companies. (Troubleshooting, offsite backups, SPAM, logs etc.)

JonWood is claiming that "yes, even small companies would want this" by making an economic argument about the value of e-mail per employee; his claim is that $4.16/month/employee, even with a handful of employees, is not that much.

Meanwhile, you seem to be making a snippy ("sorry") counter-argument that his calculations are failing to take into account that if he had that much money to spend, he would be underestimating the cost of his system administrators, which would cost more than that.

I am sorry, but I don't understand... you both seem to be agreeing with and arguing with his point, by claiming that something unrelated (the units don't even seem the same to me: money made per employee vs. money spent per administrator) was not taken into account.

Exactly, you're either small and use the free version or you're very big and $50/a/y is no big deal. There's nothing in-between. I know quite a few companies who would gladly pay something, even if the free version works for them.

I'd like the option to maybe pay $50/yr for 20 or 30 accounts. I have a couple of sites that need more than 10 accounts but can't justify paying $50 per.

But who knows? Maybe Google saw that most businesses use 15 or 20 email accounts, so if they offered something like that, they'd lose a lot of money.

Reposted from other Maps thread

Aren't the limits IP based? This page on Google's 'Geocoding Strategies' seems to say so: http://code.google.com/apis/maps/articles/geocodestrat.html#...

So as long as your call to the Maps API goes out client side, aren't you ok?

I think the Maps API is different than the Geocode in that the Maps API requires you to grab a key here:

http://code.google.com/apis/maps/signup.html

So to avoid it in that sense you'd have to force every one of your clients to go grab an API key and input it.

Look at the note at the top of the page you linked, Version 3 of the Maps API doesn't require a key.

Not only does v3 not require an API key, as of Sept 28 (a couple days before the rate limiting went into effect) sending a v2 API key to v3 will cause the request to fail:

http://groups.google.com/group/google-maps-js-api-v3-notify/...

Sounds like they're laying the ground work for v3 to use API keys again.

Here's a Tweet from the Google Maps team:

Google Maps API is still free! Just has a limit of 25k maps/day. That means pricing only affects the top 0.35% sites!

https://twitter.com/#!/GoogleMapsAPI/status/1296873737002762...

I will be switching.

I quite like the Google Maps API but in my company we're just not able to pay for things with credit cards or in fact to pay for anything we don't know the cost of up-front (i.e. variable billing). We'll happily and routinely pay hundreds of millions to suppliers so long as we have a contract, of course.

This is no doubt more of an indictment of my employer's internal processes and policies rather than Google, but it does mean that the "standard" maps platform I built will have to be re-written to use another source of tile data etc.

But it does seem like this means I have some re-work to do :-(

Is that not what Google Maps Premier is for?

http://www.google.com/enterprise/earthmaps/maps.html

I imagine so - I've already requested info.

It will take my company literally months to push through a contract for something like this. Google will be required to tender, alongside other comparable services. I just don't see that happening...

If your company is going to consume large volumes of mapcalls you will definitely need the tiered pricing which they do offer per client basis.

It's possible to get 50% discounts of the $4 price (if you do more that 100,000,000 per year)

The main problem is that my company (an executive agency of the UK government, effectively) can't enter into this kind of financial agreement easily, precisely because we have such a huge budget (numbered in the billions).

Any contract we sign needs to be for a fixed price, rather than a variable amount. And amounts above a certain limit need to go out to competitive tender, which takes a long time. (I'm not sure Google or the other players in this space would even bother to tender anyway.)

So the only option remaining is to rebuild the mapping solution using another technology to ensure continuity of service.

Surely that can't be literally true. How do you buy consumables? Surely you don't tender for contracts to provide a fixed amount of coffee or paperclips per year, choose one provider, then the amount provided just has to suffice no matter if it's too big or small?

Presuming you don't, perhaps you could find a way to get your superiors to think of things like map requests as akin to coffee or paperclips rather than a service or a piece of boxed software.

Well, if you’re playing in this big a space you could always run your own mapping server using OSM tiles. Of course, GIS guys in my experience tend to be extremely parochial and you might have difficulty in pushing that approach :/

Saying "you could always" for something like this doesn't make any sense. To undertake creation of a mapping service with no/little expertise and if the main constraint is approval timeline doesn't even seem like an option.

If it wasn't clear I meant 100mil requests, not $$$. The pricing is fixed for a tier plan. This is no different than deciding which license to acquire for a service.

Do you have any insight into how large the volume of your map requests is?

No, it's a platform I've built for multiple teams within my organisation. So far I have just two instances, but there's a chance that several of them may wish to show geographic information about the availability of healthcare services within the UK, so then numbers could get big very quickly.

They say the premiere service will be cheaper than the overage.

http://www.google.com/enterprise/earthmaps/maps-compare.html

Unfortunately, they don't list pricing in an obvious way.

10,000$ per million page views per year. So it's 10$ per 1000, more than the 8$ or 4$ per 1000 that was announced.

I don't know any big site that gets more than 10$ per 1000 pageviews with ads. It's insane!

that can't be right. the limits here get you 25k*365=9,125,000 views per year for free.

it is right... with the premier pricing.

EDIT: fill in the form you'll get the pricing... 10k$ upfront for 10MM views: http://www.google.com/support/enterprise/bin/request.py?cont...

since it's being positioned as the more cost effective option as views go up, that still doesn't make sense. Can you provide a link?

EDIT: ok, so that's an order of magnitude difference from your first post. that's well below the CPM for the "overage" charge, but well over it if you take into account the 25k initial views per day. It still doesn't make sense if they are legitimately offering that as the better high-end option.

I don't want to go all ad-hominemy, but based on your "Google Maps is dead, long live OpenStreetMap!" comment, I'm going to go ahead and take your numbers with a gigantic grain of salt :)

I am less than happy that Google Maps is dead. OpenStreetMap is still far from being comparable.

I think I made a mistake with the figures. I just checked that email again and it says 10,000$ for 1 million page views per day for public facing deployments. What a shame.

Google is changing. They used to be startup friendly. Now they focus on internal startups, and with a switch of a button cut services to potential competing startups.

I beleive it's 10K per year.

Yep http://www.google.com/enterprise/earthmaps/maps-faq.html

It starts at $10K a year.

There are some incredible Google Maps mashups, but I wonder how many make enough money to pay Google. I sure hope this isn't too tall a hurdle for PadMapper.

Yahoo maps isn't bad, and as far as I can tell they aren't enforcing their limits.

Meh.

Google maps UI is by far better than Yahoo maps. That and the street view/bike maps/etc. are features that give it a nice edge over alternatives.

Glad we (Stormpulse) made the decision to build our own mapping system.

Off topic, but your site is awesome.

Haha, awesome, thanks.

Why haven't I heard of you? Are you b2b or gov't supplier? Very nice site.

We've sold to F500, SMB, and govt/mil.

Ah, that's why. I have a slew of sites I use for Wx (just amateur- boating, interest, and so on).

Shit! There are so many sites out there using this API. And many are totally dependent on it - location is the cornerstone of their model. Zaarly, Foodspotting, AirBnb, GetAround - every location-based P2P platform.

I guess this was inevitable. But maybe we can switch to a different maps API? Google is so smooth though.

I'm really surprised that Google would make such a sudden change. I would've expected lower rates or allowing free use if AdSense ads are placed on the maps or on the same page. Such changes could've been easier to accept for many sites instead of making them look for alternatives. It appears that they decided that the branding effect of having the Google logo and copyright plastered on so many Web pages wasn't doing much.

It's time like this that RMS Free Software message really stands out. Corporations will screw you eventually no matter how sweet they sound initially.

On reading this I wondered if this was meant to target Apple? I'm sure the Maps app uses over 25k API calls a day.

Probably not. The guy who authored the blog post says that Google Maps use in Android apps won't be bound, I imagine the same would be true for iOS: https://twitter.com/thormitchell/status/129665976001236994

At $work we signed for the Premier agreement. We're paying more than $10k/year but it's still reasonable for the benefits it gives us. There are ways to keep your map views down when they aren't essential to the page. :)

When spending $12bil on a second rate hardware company is cheaper than the legal cost of patent defense, things like this will happen.

Google Maps is dead, long live OpenStreetMap!

Visual candy for the world that people didn't know they wanted - but want. +1

So, does this mean they can't monetize ads on maps?

Reply on news.ycombinator.com