> Beginning on June 11, 2018, you’ll need to enable billing with a credit card and have a valid API key for all projects. This will give you the ability to scale easily with less downtime and fewer performance issues. In addition, we’re simplifying our 18 individual APIs into three products: Maps, Routes, and Places.
> Also, in June 2016 we announced that we would stop supporting keyless usage, meaning any request that doesn’t include an API key or Client ID. This will go into effect on June 11th, and keyless access will no longer be supported. At that time, keyless calls to the Maps JavaScript API and Street View API will return low-resolution maps watermarked with “for development purposes only.” Keyless calls to any of the following APIs will return an error: Maps Static API (including Static Street View), Directions API, Distance Matrix API, Geocoding API, Geolocation API, Places API, Roads API, and Time Zone API.
The only silver lining is that www.openstreetmap.org (OSM) will hopefully get more love.
Since google admits that 98% of current users will continue to incur 0 cost this is simply a blatant way of getting a real name associated with their user data.
Any service priced $free or not tied to a credit card tends to see people play at incredibly dumb levels of abuse. Within a month of Gmail going public I remember people already trying to string accounts together as a makeshift multi-GB FTP store.
Not assuming the worst has gotten us to the current state of personal exposure to entities whose core welfare concerns does not align with or extend those of an individual.
Not assuming the worst as default will only make matters worse. Even if Google has very good and honest intentions with gathering this information at this time, you have to assume the inevitability of that changing in future, either under duress of corporate existence or being compelled by larger more powerful actors.
Have I read the above 5 years ago I would have rolled my eyes at yet another privacy nut raving on the internets.
If the worst isn't part of the strategy, they should explicitly forbid themselves from doing the worst in their Terms of Service. Or at least publicly advertised in a legally enforceable manner. The fact that they don't is pretty good red flag that they either are doing it or want to leave the door open to doing it later.
> Since google admits that 98% of current users will continue to incur 0 cost this is simply a blatant way of getting a real name associated with their user data.
I'd imagine it's more a way to make sure the remaining 2% don't just spin up a new account every time they hit a limit.
Kind of hard to compare apples to apples, there was a pretty steep minimum on previous contracts, I'd imagine that high volume users could negotiate volume discounts beyond those listed.
At least this way there's not a huge cliff between the free product and a contract for larger volumes.
It would be nice if this leads to more interest in alternatives (OSM based systems including maplet).
Previously, when google made the initial shift in this direction a few years ago, it seemed like there was a brief period when more people were considering alternatives, but since most people were able to continue using the free tier this tapered off pretty quickly.
Leaflet is a nice javascript library for slippy maps, but you still need to bring your own tileserver (OSM is not free to embed for anything beyond light usage I believe)
Are all those loads needed? In other words, how often do people interact with them? (Figuring how often they look at them is even more effort.) This kind of changes usually are made to discourage inefficient use of the API, which could be delayed until specific events happen: scrolling, mouseover, timer, etc. I'm not saying that's necessarily your case.
I can probably lower the loads somewhat, but I use other maps services too (street view, etc.) and this $200 free credit seems like for all of the services combined monthly, so one does not get a $200 free credit for each service and costs will quickly add up.
I see that the limit you mention is for dynamic maps. Somewhere else, they say "Requests for Dynamic Maps and Dynamic Street View (Street View Panoramas) will be billed at a higher rate than requests for Static Maps and Static Street View. Dynamic Maps and Street View images respond to user input; the Static versions do not."
I imagine that this is an incentive for developers to use static maps (which you could replace with dynamic ones when the user clicks or hovers on them). If you used static maps alone, that's 300K loads per month, or $480. You would deduct the $200 credit, for a $280 bill, but then pay full price for all the other services. I can see how that adds up quickly...
I did some comparisons with the old pricing as well. For geocoding it seems the free tier got halved (assuming average usage over a month) while the paid tier got x8 more expensive.
Not a native English speaker and also not that well versed when it comes to marketing talk, but what exactly does "$200 monthly free credit" mean? (from this page: https://cloud.google.com/maps-platform/user-guide/pricing-ch...). Do you get a sort of voucher with which you'll be able to use Google Maps services worth $200 each month? Will Google automatically start billing you once you surpass that quota? I'm in the dark here.
Later edit: I've found the answer to my first question, as in yes, the "$200 monthly free credit" acts like a voucher, sort of:
> Can I still use Google Maps Platform for free? Yes. Starting June 11, 2018, when you enable billing, you get $200 free usage every month for Maps, Routes, or Places. Based on the millions of users using our APIs today, around 98% of them can continue to use Google Maps Platform for free with this credit.
Still don't know the details of how to set up things so that I'll never surpass that traffic quota of theirs.
I'm part of that 98% they mention with a couple of small personal websites that aren't intended to bring any revenue, I'm not that hyped with giving Google my billing details.
Later later edit:
It seems that it's possible not to give them your billing details so that they'll automatically "degrade" your API calls once the traffic quota is reached without billing you. I guess that's the route I will take, it's the least amount of effort and seeing as my projects won't reach those quotas it also makes sense not to worry about API calls "degradation".
> Is a credit card or billing account required?
Even though your first $200 of monthly usage is free, all Google Cloud Platform services require a credit card and billing account, to cover any amount you spend over this free credit. If you are billed, we'll credit your account for the first $200 of monthly usage. If you choose not to add a billing account, there is a risk that if your usage exceeds $200 in a given month, your Maps API implementation will be degraded or other API requests will return an error. If your estimated usage will be above $200 a month and you don't have a credit or debit card to set up a billing account, a local Google Maps Partner may be able to help.
Not neccessarily obiovus, because the current quotas are also per project. So if you get a 1000 calls a day then every different map projects in your acocunt gets 1000 calls. This suggests the new free tier can also be per project.
"All Google Cloud Platform services require a credit card and billing account... If you choose not to add a billing account..."
So is it not actually required? Or is maps not considered a "Google Cloud Platform service"?
I have no idea what's going on with the pricing page. What's the difference between "embed" and "embed advanced" maps? (other than 1.4¢ per load). I'm assuming that 'native' means mobile, but I'm not really sure. I'm currently using the JS API, which I think is equivalent to 'dynamic maps'.
It's a shame that a product that probably took 7- or 8-figure dollars worth of development effort, didn't have some beta testers do a once-over on the documentation.
I wish it were an isolated instance. For one google storage product we found 2 pricing pages with 2 different billing models and a Google rep who added a 3rd billing model into the mix.
Billing models have become a new competitive frontier and lead funnel strategy in Saas. All but the most cursory usage has become "Enterprise" and begets a sales call.
> Is a credit card or billing account required? Even though your first $200 of monthly usage is free, all Google Cloud Platform services require a credit card and billing account, to cover any amount you spend over this free credit. If you are billed, we'll credit your account for the first $200 of monthly usage. If you choose not to add a billing account, there is a risk that if your usage exceeds $200 in a given month, your Maps API implementation will be degraded or other API requests will return an error. If your estimated usage will be above $200 a month and you don't have a credit or debit card to set up a billing account, a local Google Maps Partner may be able to help.
It's still not clear whether you have to give them your CC details or not. I once made this mistake and gave Google my CC details for Adwords, and I regretted it. I found the hard way they will run the ads for quite some time after I stopped the ads and charged me. Of course it was difficult to contact a human and ever if I managed to do it, it wouldn't help. One experience smarter and I only use bank transfer, thereby setting a true hard limit and making sure a stop means a stop.
From the email I received it states you must have a billing account to qualify for the $200 Free otherwise you won't be able to use it.
"In addition, this change will require you to enable billing and associate it with all of your Google Maps Platform projects. Creating a billing account helps us better understand your usage so we can continue developing helpful products."
If you don't add a billing account it:
"will return low-resolution maps watermarked with “for development purposes only.”"
Does anyone know why Amazon doesn't have a maps offering on AWS? I've heard their internal logistics GIS services are pretty advanced and it surprises me they don't have a GIS category on AWS to the point I'm convinced there must some very specific reason they've chosen not to implement. Companies like MapBox and Carto seem to be doing well, Microsoft and Google have map services, demand seems to exist. Additionally it seems they would find value in possessing a list of companies utilizing such services and be able to offer some interesting services. It's puzzling.
AWS seems to me to mostly be a way to configure and sell capacity on hardware. Amazon probably has a lot of fantastic data and software that isn't available outside the company.
Why don't they try to personalize or target the ads on my Kindle? It would be nice to not see ads for books I've already purchased from Amazon. They inscrutably leave a lot of missed opportunities on the table.
"The Amazon Maps API is a programming interface that enables you to add mapping capabilities to your app on Fire phone and most Fire tablets."
"The Amazon Maps API v2 is compatible with Fire OS 4.5.1 and higher, which is available on 3rd and 4th generation Fire tablets, and Fire phone. The Amazon Maps API v1 is compatible with the Fire phone and all Fire tablets except Kindle Fire (1st Generation)."
Did not know about this, thanks for sharing. It almost increases the mystery in my mind. It appears this API is only available for use in Amazon Android apps, it's certainly marketed that way. It's as if they went 99% of the way of having a full map offering, but decided that Maps were somehow going to drive Kindle Fire adoption. Very interesting.
More likely than not, because the data is not theirs (I think it's Navteq->Nokia->Here), they wouldn't be able to resell it with a markup at a competitive price, so they don't even try.
Supply is expensive, and running a 'maps offering' is more of a data-collection and data-maintenance challenge than a logistical challenge. Of course AWS probably will offer such a service eventually, simply to complete their Gotta Catch Them All product strategy, but for now they've seemed to focus on things where their value is in software and infrastructure, not in having a 500 person staff of people with the background and skills to validate and compile geospatial data.
To the inevitable takes on this, no, it's not just a matter of downloading OpenStreetMap or any other data source.
> From some basic OSM v Google tests, OSM coverage was much lower for things like place names or speed limits.
In which region did you test? The local mapping community can make a night and day difference.
In Africa there are places almost in unmapped on Google that are extensively mapped on OSM (Google doesn't give a damn if there isn't money to be made). In western Europe it's pretty good, especially in the Netherlands and Germany (that I know of at least, perhaps other countries too). America, from what I've seen (I've never been there so never needed to use it) it can be either reasonably comprehensive or a total mess depending on the city or state or something.
> it can be either reasonably comprehensive or a total mess depending on the city or state or something.
I'm currently evaluating US coverage and this has been my experience, though even cases with better coverage it still fell short of Google Maps.
It's a tough problem in the sense that people have generally gotten used to the excellent coverage of Google Maps, so anything deviating in even a minor way is seen as noticeably inferior.
I'd love to know if others are running into similar issues in the US.
If you mean business locations and the like, OSM is more or less empty in the US. Cities and towns and similar should be okay.
What it amounts to is that there aren't any openly licensed business geodatabases and there aren't really that many people working on OSM in the US, especially commercial POIs (where people buy data from Yelp or Foursquare or Factual or whatever instead).
This is the nth time answering this, but, well: Updates. Doing them efficiently. Vandalism detection. Geocoding and directions. Devops: this is not a trivial computing challenge. The list goes on.
They don't usually share data across the groups like that.
They would essentially be asking Amazon retail to share proprietary data with AWS, which would then allow someone like FedEx to start using AWS to get that proprietary data.
That data is probably much more valuable to Amazon retail hidden then it would ever be to AWS exposed.
They invested in and are partnering with Here for maps and navigation. Having worked there, providing global maps and navigation is a very non trivial thing to pull off; which is why there are only handful of companies doing their own maps. It took Apple and Google many years to stop depending on third party maps that they were licensing from Teleatlas (now owned by Tom Tom) and Navteq (now part of Here), which were at the time the only suppliers of maps and navigation world wide. Apple is still in the process of shifting away from Tom Tom I think. Ironically both got more serious about doing their own maps after Nokia bought Navteq.
So, Amazon did the smart thing and simply partnered instead of reinventing that wheel so they could have a maps solution for their Android fork and not depend on Google Maps while still implementing the relevant APIs for app compatibility. Here is in fact using AWS so that ticks that box ...
OSM is nice but entrenched in licensing issues that are preventing adoption by big corporations. That's the reason that you'll be hard pressed to find anything but niche products for navigation based on OSM. It's a great solution if you are not mixing your own data sources with what OSM provides but that is not an option for most commercial players and OSM licensing is all about preventing exactly that. Microsoft had OSM maps for a while on Bing but they couldn't do any fancy features based on their own data on top of it so it got scrapped.
I'm just starting a side project that uses similar, i think it's the pricing around autocomplete and places information that seems like the most painful part to me.
Currently I'm using it to search and add points on a map, it seems like it could really add up
I was trying to set up my first gmaps project a couple of weeks ago using node and ran into a bunch of issues that were hard to resolve due to lack up updated documentation.
Switched over to mapbox and got everything working in like 5 min or so.
+1
TallyGo is $2.50/1000 routes with our mapping provided for free as long as you're using routes. We'd love to have you check it out: https://www.tallygo.com/pricing
What services do you use? I'm working on a startup in this space, and I could definitely recommend a set of services that replace Google for much cheaper (even if you didn't use us ;) ).
Primarily just maps; real estate sector. The $20k is a ballpark for one project, another would add another ~$5k a month.
It's a good motivation to improve our usage of their APIs and I'm sure that I can bring down those cost estimates (by reducing API usage overall and by switching some to use the static map API - although the caps on resolution gave us problems with that in the past).
That makes sense; well, if you wish, please reach out to me at luke@stadiamaps.com
We have a lot of customers in real estate, and would love to set you up with something that would be significantly less than $25k. These situations are why we started the company. :)
I think our tiers work very well for the real estate space overall.
I'll have a look, thanks. Your pricing looks much more reasonable - I suspect you're going to be seeing a large spike in demand.
Initial work on reducing load looks very promising though (lazy loading only when users scroll as far as the map on property pages, pooling map instances/dom elements and reusing them across a React app so that they're not recreated) - both things I really should have been doing already anyway. I guess that's part of the motivation for the change - even if I don't pay anything, I'm now costing Google less.
and you have 30 days to somehow make it work or switch. for an API service that's seems pretty irresponsible to it's users. You would think they'd give you a few months notice at minimum.
I'm an ex-Mapzen teammember, and we've set up a new, paid, geocoding service called https://geocode.earth. I suspect we can give you what you need around geocoding, especially autocomplete, for much less than Google. Feel free to request an invite or reach out to me personally.
This won't happen unless you have a huge traffic which usually should translate to revenues that can support it. There is nothing wrong with services that you use to generate profits but have to pay for. Services like maps with 99% coverage of the world takes enormous amount of efforts over many years of sweat by very talented team. I know of countries where government wasn't able to produce good maps for decades but Google was. Currently while I travel, I very heavily rely on Google Maps. It's a pretty amazing service and I have paid more for stupid Angry Bird.
100 users hitting 10 pages a day = 1000 map loads a day. 1000 map loads a day * 30 days = 30000 map loads which is over the "free" amount. This assumes NO api calls at all, only map loads.
This change will affect all apps that have any sort of actual user base. I think you may be underestimating the how quickly you can go through API calls even without many users.
While I understand they want to make money, giving developers 30 days to now pay 40x what they paid before is ridiculous.
> This won't happen unless you have a huge traffic which usually should translate to revenues that can support it.
For a B2B product where every user/visitor pays a high fee, sure. However, for a consumer website that is either ad-supported, or where a small percentage of visitors pays a low fee, this is definitely not true. The new pricing is insanely expensive.
I'm putting together statements from those affected by the huge price increase to send to the Google Maps product team as a joint message. If you'd like to add yours, please email me at nick@instantstreetview.com
I wish it would be easier for people to contribute to open source projects (like wiki voyage etc) than to Google Places. (it's a bit off topic and I like the Places API, but that's a huge amount of valuable data that contributors are submitting there for free)
Agreed. I contribute to Google maps a lot and I've looked into contributing to open street maps but the options seemed pretty clunky. I recommend mapbox to all of my clients over Google maps.
Definitely strong with map visualization. They're based on OSM map data, so weaker on anything that as to do with geocoding and routing (business data, speed data, traffic data). For good routing and search, you need to look at professional platforms from TomTom Maps, HERE, ESRI or similar.
I think maps.me is one of the few apps that are easy to use and contribute to OSM. Unfortunately there was some controversy, because maps got vandalized or filled with trivial/garbage data, because user didn't understand that their notes ended up in a public database.
Did they change the usage limits of the Distance Matrix API? Previously they limited to "Maximum of 25 origins and 25 destinations per server-side request", even if you were a premium package subscriber, which is quite restrictive...
> If you are currently using an API key, please read the email we sent you on May 1st, which provides a high-level estimate of whether your future bill will change.
Did anyone actually get this mail? I am using a key and even got a receipt yesterday but did not get this mail about pricing changes (user from Germany).
Edit: Wow. I am currently using Places Radarsearch for a side project which currently has a (generous) free limit of 150.000 calls per day. If I read the pricing table correctly, I am now getting 5000 calls per month on the free tier (which also has to suffice for all other maps API calls). That's a 900-fold reduction..
I suspect that's 98% of users in the literal sense, with a very long tail of them having very little traffic. It's not the same as 98% of user traffic, i.e. API requests. :-)
If you were using the free tier of Places Nearby or Text Searches (aka radar) and staying just under the daily search limit, your previous yearly bill of $0 from Google will now be roughly:
As someone who's only every used Google Maps, can anyone suggest a solid alternative? Geocoding and directions would be the minimum I'd need - with solid data in Australia.
What kind of places details are you looking for?
Depending on the geographic region, Mapbox has data for regions, cities, addresses, zipcodes, and points-of-interest (POI).
How long does it take for an edit to propagate from OSM to Mapbox's tiles? Average/worst case. I know this is a tough problem, so I don't envy you. :-)
Well, they probably have some kind of tile "lint", need to precompute a bunch of routing data (depending on what their actual algorithm is), sync all their rendering/prerendering pipelines at a common revision, propagate outputs to cache servers, etc. So it can't easily be done at every change, unless they have a design where small incremental deltas are explicitly supported by all their services. It is a time- and resource-intensive process. Still, three months sounds like a lot.
The concern seems to be moderation (but they haven't made a clear statement about it either). The current pace of updates is slower than what they were doing a year ago, so it probably isn't a technical thing.
Their routing engine updates more frequently than the rendering data.
OSM sub-10 minute updates really depends on zoom level and data. With low zooms the update can take weeks even there, as I recently learned in a hard way. It is more important to have consistent and laundered map data than someone's last minute changes.
As you correctly state, their servers aren’t free and hammering them will incur a ban. On the other hand if you are a high profile user why not copy the free data from http://planet.openstreetmap.org and have your own server? Or if this is technically difficult, you can find paid for services with OSM data (I guess if you go down this road, it’s like paying Google)
One thing I like about AWS is that all services have a pricing section. I am interested in the routing service (https://cloud.google.com/maps-platform/routes/) but there is only a contact sales button, no public pricing info available.
I find it very interesting that Google Maps has 1 billion Monthly Active Users compared to Facebook’s 2+ billion. Where is this huge difference coming from?
I lot of people never change their travel patterns. Then there are those who use Waze (I think these are kept separate even though owned by Google) or Apple Maps.
Probably from people that don't use maps all that often? Back before I lived in the bay area I rarely used maps simply because I didn't need them. The area was small enough that I just knew where everything was.
What is "embed advanced"? A search turns up nothing. I'd hate to make a custom map style, embed a map with it, and suddenly be charged upwards of $14 per thousand calls. Seems pretty steep regardless of what it is. Mapbox doesn't charge that much for anything on a per-call basis.
iOS11 added a "Desktop view" mode but I am not sure the feature is well known (Android has had the feature forever, so by mobile, you mean iOS I think).
From memory, in Safari, tap share, then scroll across to desktop view. I don't have an iDevice with me at moment, but it should work.
By "mobile" I mean Android. And since Android has many browsers "Desktop view" may or may not exist, and if it does exist and can mean many different things. I don't bother to use it because I never found it to work properly. I haven't tried for a couple of years, though.
However I have found that it is quite unreliable at working. Not sure if it's just changing the viewport for CSS, or the user-agent, or what. But it frequently doesn't work.
Ive always wondered why Google Maps hasnttaken that of a topological yellow pages - where you have a limited blurb that can appear on your location beacon if your a business and if I search for "thing near me" your location and blurb can pop on a users' map and if they click on you to go to you, think of it as an ad click to your location.
Or is this already a thing?
EDIT: This looks exactly like what they are doing with that little rating icon above the building...
I think it's one of the things that was behind the big push for Google+. Facebook won, but at the time they hadn't quite yet become the default place for small businesses to do social media.
Mapbox has an excellent alternative map platform based largely but not entirely on OpenStreetMap data. Mapbox GL JS in particular is a joy to work with and extend. https://www.mapbox.com/
It's not free, but it's very easy to use and the pricing is roughly 1/10th of Google's when you get past the free tier: https://www.mapbox.com/pricing/
Mapzen used to be an alternative but sadly they shut down. However many of their products live on in open source or paid-hosted alternatives. See https://www.nextzen.org/ for a start.
Mapbox is pretty expensive for small commercial apps, $499 flat rate + usage costs. If your commercial usage is on the smaller side, the new Google Maps pricing looks really good.
I was a Mapzen customer, their pricing model was great, very sorry to see it go.
I've never understood Mapbox pricing. $499 lets you give maps access to 250 users. That's steep for solo devs / small shops that want to create a map-based app. And there's no way of knowing how that pricing will scale without contacting them for an enterprise plan!
Hey Tyler, I work as a customer success manager for our commercial segment and are happy to say we're willing to work with you on a price that fits your business. If you write in through our contact forms just ask to speak with Erin and we can work something out! https://www.mapbox.com/contact/sales/
I'm in the process of comparing Mapbox with Google Maps Platform. The usage would be for large full-screen maps with lots of interaction. Comparing map loads/views is not apples to apples. Google counts a single map "load" similar to a page view [1]. The number of tiles loaded doesn't influence cost. With Mapbox, a map "view" is 4-15 tiles (depending on how they're styled) [2].
OsmAnd can do this. It's generally recommended as the OpenStreetMap app, but almost everyone uses it with offline maps. If you click past that and enable the built-in online tiles addon, you can have tiles. Search and navigation, I don't remember seeing it in recent versions but there are so many options, I'm almost certain it must be in there somewhere. Search for custom or online navigation provider in OsmAnd and you should be able to find it.
As you see, it's not exactly built for this type of usage, but it is supported and the app is generally considered as the best alternative out there, so it might be worth a try!
For good search and routing, you need a high-quality map data designed to deliver professional find, routing and navigation services. I recommend you take a look at developer.tomtommaps.com
I'm working on the design of a system right now that is intended to be commercial if we can bootstrap it with profits. I had already decided to find an alternative to Google Maps because of the risk of getting technologically locked into a completely custom API/service. What, for example, would happen to our profits and bootstrapping plans if Google decided to, say, double or triple its pricing overnight? Big trouble right?
So, here we have people who built around Google Maps wishing the prices were only tripled. I'm not yet sure what we will do, but we'll try to improvise something from less-featureful maps before we'll build a nest out on a Google branch and dare them to cut it off. I'm glad I learned my lesson from the experiences of others before this even happened. This is a great reminder about the hazards of designing around single-source, critical inputs.
I'll tag on to Nelson's post to plug a new geocoding service two of us ex-Mapzen folks started called https://geocode.earth.
We offer geocoding, especially autocomplete, with much more liberal usage restrictions than either Mapquest or Google, and some pretty powerful features like address interpolation.
As we're still really getting the platform around the geocoder off the ground, we don't have a permanent free tier, but I'd be happy to set anyone up with a demo key if they are looking for something new.
Also TomTom offers an aternative for Google Maps with high quality maps, geocoding, routing and traffic APIs + web and mobile SDK: https://developer.tomtom.com/tomtom-maps-apis-developers Pricing includes a free tier of 2,500 transactions/day and is much more affordable than the new Google Maps prices. I work for TomTom so in case you have any questions, feel free to contact me at developerrelations@tomtom.com.
Mapbox is not cheap either.
I use geocode.xyz to batch geocode data for Singapore and it works great (there is a free batch geocoding api)
It even correctly geocodes some addresses I could not find on Google Maps. And you can get unlimited access for $100 per month.
TallyGo is a startup offering a routing SDK for Android & iOS. The pricing is much more affordable and capability for app integration is strong. It has realtime traffic & advanced routing. (Disclaimer: I work as a developer at TallyGo)
I don’t know how competitive the pricing is, but it is interesting how the comments go when Google does try and charge a high price for something over simply monetizing via advertising.
I wonder if the perception would be different if Google Maps were a standalone startup developing a world class mapping product. Where the common manta would be along the lines of double your pricing.
I think the problem is users don't want to pay and have their data scrutinized for ads. It would be like if you paid Facebook and still had your data going to many third parties.
I think Mapbox has a very different reputation for this reason (also they contribute back to OpenStreetMap).
Google doesn't store user data when you use the paid version of maps. For example when you call places API on the premium plan nothing is ever stored on Google's side after the query is made and results returned.
Having contacted Google Maps sales years ago, their salespeople were some of the worst I've ever encountered--rude, inconsistent aggressive pricing, and vaguely threatening. I avoid them at all costs, and as others mention, Mapbox is a great alternative.
Back when I was working on a map heavy app, Google Maps API seemed designed to discourage usage. Bad/opaque pricing, usage limits, assumption that Google robots will pull the plug with no warning. In fact all Google APIs seem to be this way - they are not really interested in enabling 3rd party apps through their APIs. Seems bizarre to me, because they could gobble up the businesses of companies like Mapbox overnight if they made an effort here.
Maybe I'm missing it, but I'm not seeing what the cost is for higher resolution maps. As I understand it, in addition to usage beyond what the free account allowed, a paid account used to also give you access to higher resolution map tiles and a higher zoom level via their API.
So all those sites have 30 days to figure out something or start shelling out money? That's really short notice
From what I can find the "dashboard" only shows the highly variable "per day" numbers so you can't even see how many calls you're doing per month... which is pretty shady
For a company that touts "do no evil" they are going to hurt a lot of free websites and small businesses / projects with basically no notice.
Wow. I teach a course of mobile programming with Android. Now I will have to remove the section explaining Google maps API as I cannot ask my students to add their Credit card info online just for the course... Any other free alternative?
Fwiw, there is a viable work around. Android (and iOS) usage of the Google Maps APIs remain free. See the Mobile Native Dynamic Maps line on our pricing sheet: https://cloud.google.com/maps-platform/pricing/sheet/
You can construct API keys that are restricted to an Android signing key, use it for the duration of the class, and then revoke it after the course work is done.
I'm an ex-Mapzen alum who runs https://geocode.earth. We don't have a permanent free tier at the moment (our geocoder is solid but the service is new!), but we want to find ways to contribute to the community where we can. I would be happy to set up a temporary free tier on our service for one of your classes.
Shoot me an email, julian at geocode dot earth and we can figure something out
I just wanted to add that geocod.io is now a viable alternative to Google Maps for geocoding in terms of both accuracy and pricing.
We’re going to continue having a no-credit-card-required 2,500/day free limit, as well as the most affordable pricing, without restrictions. Including a plan that allows for unlimited geocoding. https://geocod.io/pricing
Full disclaimer: I’m one of the Geocodio founders.
Wow, this is a massive price increase... free tier cut from 750,000 dynamic map loads per month to 56,571! With new prices it'd cost $4,023/mo to pay for what used to be free. Another way of putting it: used to get 25,000 free dynamic map loads per day, now only get 1,886 free map loads per day. 92% less for free. And this is assuming you use the full $200 credit only for the dynamic map loads and not partially for the other Google Maps APIs.
Gah, this is so confusing, I've got small sites developed for clients many many years old running various maps based stuff off API Keys in my account, now I have to add MY billing information to them or get the client to set themselves up a google cloud platform account and developer and project and keys and ......
Good grief, that's totally impractical to have website administrators, my clients, to do that, even I as a seasoned developer hate trying to navigate around the complex labyrinth that is the Google developer console stuff.
It's all so complicated. I just want a page I can point my customer to, "go here, type in your domain name, enter your credit card number in case you need to be billed by Google for usage, and tell me the API key it spits out".
On top of that, it's still of course totally open to abuse and quota theft, referer checking is not security. I can see a burgeoning black market in api keys forming.
We support cases like this by using domain names as authentication.
We plan to eventually support automatic spinning off of domains into their own account (e.g., for your clients after you're finished with a project). But for now, they could set up their own account with their domain and it would just work for you as a dev.
It is totally hidden but you don’t actually have to add a billing account if you’re sure that you stay below the $200 monthly usage.
You can find it here: https://cloud.google.com/maps-platform/user-guide/pricing-ch... "Is a credit card or billing account required?"
I also experienced similar issues when I was still freelancing and managed the hosting for some clients.
I wanted to host clients on Google Cloud or AWS but letting clients make an account they are unable to manage looked like a bad idea to me. If their website gets a lot of unexpected (maybe unwanted) traffic, they get a large bill which they blame me for. Because of this reason I had to go with shared hosting for small clients, and Heroku for companies that cared about uptime so everyone always knew what the costs are.
Something like billing limits would fix this. So freelancers could host their clients websites on their own account (or using AWS organizations) and charge their clients the billing limit for a year upfront.
I run https://onthegomap.com which uses Google Maps API. Last month it cost about $500 (for direction requests). Under the new pricing model it will cost $30,000 per month: $2,500 for page views, $27,000 for direction requests, and $1,000 for elevation request. It makes $300-500 per month in ad revenue. Time to shut down?
I don't think that's specific enough. For example all the places API calls are grouped together but there are 13 types at different prices. I don't see any way to figure out how much we'll be paying.
For me looks like I went from basically free to $600/mo just for map views. I have no idea how much more I'll have to pay for API calls because they completely obfuscate the data.
Now we have 30 days to figure out how to make the money work or find another solution. Pretty crappy of an API provider (let alone Google) to do this. At least now we know not to trust our apps / sites infrastructure to Google... they aren't the same "do no evil" company they used to be.
You can host your own OSM-based mapping for a fraction of your current cost, or move to Mapbox or similar. Drop in on https://help.osm.org/ or the #osm IRC channel on OFTC if you need any help.
Thunderforest is a great start for tiles. I'm not aware of a discrete elevation provider, though Graphhopper can be made elevation-aware, I believe. Place search is more difficult (if you're thinking POIs) and it might be worth investigating Foursquare etc.
What is the simplest way to self host a map that will serve as bakgrund to WMS data in Leaflet? I need nothing fancy, can do without deep zoom. The usage is within an app so embedding in the app could be an option. I find the mapping landscape confusing.
With the new pricing I think anyone who has been paying any significant amount for Google Maps usage would be better served with MapBox. The free tier usage in GMaps was rather lenient, but with this pricing change I don't see many reasons to stick with Google Maps's subpar developer experience and performance (except for the extra churn required).
Open source idea: Google Maps -> MapBox shim layer
What fresh hell is this transition tool? It goes through a few weird, semi/non-styled authentication pages on a variety of domains (e.g. withgoogle.com). It uses non-standard animations that look broken. The whole things is a disaster that is clearly not finished.
The whole thing smacks of 90-00's era Microsoft 'wizards'.
If i did my math correctly, then in my situation 0$ converted to ~400$. Google gave 500$ free credit, which will run out in 1 month. Simplest way to optimize things will be to replace dynamic maps with static and remove StreetView. Then i should be ~200$ monthly limit. Which is not safe place to be.
Edit: Actually 500$ will last for 3 months. As each month there are 200$ free. Thus if bill is 400$, then half if covered by the free tier and second half from credit.
There's (justifiably) a lot of love for Mapbox in this thread as an alternative, but as remarked, their pricing isn't particularly cheap (why should it be, they sell on the quality of the product).
So the other option is to self-host tools using OpenStreetMap data. Some pointers:
* Raster tiled maps: the canonical instructions are at switch2osm.org; put it on commodity hardware such as Hetzner/OVH.
* Vector maps: a lot of development in this area at the moment and you could happily lose yourself for days among all the emerging solutions, but probably the easiest out-of-the-box solution is openmaptiles.org.
* Geocoding: Photon (https://github.com/komoot/photon) is the easiest to set up because it comes with precompiled indexes - no need to generate your own. Based on ElasticSearch, needs SSD for reasonable performance.
* Routing: OSRM (Mapbox's routing engine) and Graphhopper are both self-hostable. OSRM is very RAM-hungry, but blazingly fast.
This is just a start - there are many more options available, not least the whole stack developed by Mapzen before they shut down. The raw OSM data is downloadable from planet.openstreetmap.org (full dump) or http://download.geofabrik.de/ (regional extracts).
I recently did an evaluation of Valhalla, ORSM, and Graphhopper for routing based on openstreetmap data.
Valhalla was nice in that it could preprocess and serve results with only a few GB of RAM, but it could only handle <1rps as the end was more than a few miles from the start since it doesn’t use contraction hierarchies.
OSRM was super fast, especially for very short routes (>1k rps on a single core) but it requires 300+GB of RAM to preprocess the world and 64GB of RAM to serve direction requests which gets a bit too expensive for me.
Graphhopper was nearly as fast as OSRM for short routes, but just as fast for longer routes. It can preprocess the world with only 64GB of RAM, and can serve direction requests happily on machine with only 16GB of RAM.
I ended up going with Graphhopper, but still may need to shut down my site...
For resources to preprocess the world I ran Valhalla on my laptop and it preprocessed the world overnight. For graphhopper I kept running it on bigger and bigger digitalocean droplets until it just barely finished on their $640/month 64gb cpu-optimized droplet in about 12h. I didn’t even try osrm after reading that it needed about 300gb of ram.
If you enter your credit card in the API Console, you are enabling Google to bill you for any other API you use which you were relying on the free quota, so be careful. Happened to me.
For those who are not in 98% that Google believe will not be affected, I combined a list of alternative to Google Maps API, with focus on geocoding - https://goo.gl/5GUVNd
Not affiliated with any of the providers listed and mainly based on this thread, just added there pricing information, free tier limits and reference.
1. To avoid making a loss due to map costs, sites may have to create a source of revenue which scales in proportion to the amount of maps used. Advertising? Google Advertising?
2. It seems quite easy to attack someone by linking to their map content on social media. Loads of traffic = loads of cost. So sites may need to consider devising a 'mapping cost defense' of some sort.
I don't think 2 would work, we use static maps and they only work based on the site they're loaded from. They won't even work on our dev instance because I've not whitelisted the domain.
Unless lots of people are going to your actual site, of course.
I have a project which runs a variety of sites for our network - currently ~900 different domains. Restricting keys by domain name is not really practical for us.
Many do, but not all, and even then just wildcarding the common bits would leave it pretty much wide open. Maybe if they allowed regular expressions then it could be targeted reasonably closely, or if there was an API I could use to generate the keys then I could provision specific credentials for each site...
Well a truly bizarre thing happened, the Mapfit staff just altered the text on their pricing box from 100,000/day to 2,500/day (Google cache confirms). Doesn't look very professional.
What they don't tell you is the underlying data is of extremely dubious quality. It's a hodgepodge of proprietary data, data automatically generated from satellite imagery and open government data from various sources all merged together in an automatic or at least thoroughly non-expert way. To a cartography geek it's an absolute nightmare.
I think a lot of people are using the stylized Map views over the satellite views. I’m my experience, those are pretty good for showing where things are, maybe if not topographically accurate.
You've got a point. If it's just about showing things on a map, OSM will do. When it comes to routing or geocoding, you need quality, consistency and accuracy. And that is only offered by map makers like Here or TomTom. Both run their own Maps API platforms.
I run https://www.instantstreetview.com which does 100,000+ 'map loads' per day. The price for these has effectively increased by a factor of 28 ($0.50 -> $14 per 1000), which is completely unsustainable ($1,400 daily!)
Will be extremely sad to see fun maps projects disappear as a result of this. We're also seeing a huge increase in cost (we are a paying user already) at https://www.openrent.co.uk/, but I kind of hoped commercial projects would be treated differently to non-commercial ones.
When I spoke to the sales team last year, they admitted they had no appropriate plan to suit small, but popular projects.
They attempted to shoehorn me onto the premium plan but many aspects of it are geared toward larger entities - very steep initial cost (per map load) with discounts for huge volumes.
So I signed up to Mapbox, I wonder how the free tier works. I didn't enter any billing details, so what will happen if I reach the 50,000 calls within a month, will it stop working or will I be invoiced later for the calls over 50,000?
We'll keep your maps on and send you a note letting you know you're over the limit. At the end of your billing cycle we'll invoice you for anything you use beyond the limit and, once the invoice is out, you'll have a ten day grace period to add your card.
Something is not totally clear and will make a big difference. Prior to this with standard it was 25,000 free loads per day. Is this being dropped? It seems like it but I received an email saying based on my usage I would be within the $200 free credit limit.
I can provide some of you with free unlimited mapdata solutions for geocoding and base layer data with accreditation. I run my own OSM server instances. Hit support on ScribbleMaps.com and ask for Jonathan so the ticket is forwarded to me.
It seems like sticking a CDN in front of the static map API could be used to reduce costs quite a bit there as well, especially if you forced it to cache more aggressively than the cache control headers returned by Google indicate. Has anyone tried this?
Technically possible, but sadly prohibited by their T&Cs[1]. It's a bit annoying, if they wanted to charge per impression they should charge by impression rather than having per-api-call pricing but forcing you to call the API repeatedly for the same thing.
Google can change their prices at any time, and they just proved that they do.
There is some kind of trust I have/had in cloud providers to be fair when it comes down to vendor lock-in. One of the things I expect a cloud provider to do is not suddenly increase prices and instead give their clients a reasonable period (a year would be perfect, but I think 6 months is a reasonable period) so they can change their usage to minimize their new costs according to the new pricing or transfer to another provider if the costs are too high.
But Google just proved that they don’t really care about their clients.
If you made an app that heavily relied on the Google Maps API, and you spend $3K a month now, you just received a mail that your costs might be increasing to $60K a month and you might need to drop all your current projects/tasks to migrate the app to another service.
Larger size (and higher resolution) static images are currently only available to paid up "premium" map subscribers; presumably this functionality will now become available to everyone, as we'll all have to be paid users (even if you don't end up paying anything because of the $200 in credit). Is there any confirmation of this anywhere? Larger sizes make switching to static maps more viable for us (one reason we switched _to_ the dynamic maps for some use cases in the first place).
reply