How do you convince Scrooge McDuck to part with money to pay API licensing costs to Google or Microsoft?
The other problem is that if I’ve got to spend that kind of money, I’m going to be developing larger internal applications than I would externally. You can’t do a simple internal mashup if you have to pay large enterprise licensing costs. I was told by one potential client who wanted a Virtual Earth application that “For that price we might as well buy another ArcGIS Server license”. For someone to say that it is cheaper to roll out ArcGIS Server for a simple mashup really puts it all into perspective and if you are going to develop with ArcGIS Server, you’ll probably have a larger application than a “simple” Virtual Earth map.
So I’m pleading with Microsoft and Google to revisit their map API licensing and move toward more of an Amazon AWS type service that will help get their mapping tools into more places than today. Everyone wants a Google or Microsoft map on their intranet website, but the current licensing is killing projects before they can start.
The GeoMonkey is a huge supporter of Amazon Web Services
39 Comments
Probably licensing restrictions imposed by external geodata suppliers, of which Google/MS have many. Trying to manage/placate/track licensing/pay them all/resolve license violations must be a nightmare.
Thus Google moving towards user-created (royalty-free) topographic data, Street View (will have generated GPS road data in direct competition with geodata suppliers) and directly funding satellite imagery collectors such as Global Eye. MS has gone the aerial photography route.
I might have bought that argument a couple years ago, but both Google and Microsoft have renegotiated their providers in the past year and they should be able to provide pay as you go.
One lump sum payment or micro payments, the hassle for them should be the same. We use OSM and OpenLayers internally because of the licensing limitations. Users would rather have nice aerials, but when they learn the costs, OSM is perfect.
Amen.
We can preach the virtues of simple, rapidly developed lightweight apps until the cows come home, but the prerequisite of an $8-$10K annual licensing cost before any custom development is even spoken of, and a lot of potential clients are feeling less than”agile”.
And yes, we know, that for a large enterprise that already recognizes the importance of “location”, the licensing fees aren’t a show-stopper. But the much bigger market of enterprises with little experience with geographic visualization and spatial analysis–but lots of experience with failed IT projects–will count the licensing set-up as a red flag.
I’ll end by saying if I were a big company with lots of cash, had a robust web mapping platform that has only captured a single-digit percentage of the mash-up market, AND I was releasing a flagship database product with new spatial capabilities, well, I might look at this as a great opportunity for grabbing business-mapping market share by making the barriers-to-entry for in-house map development as low as possible.
Brian
p.s. I’ve been especially not enjoying explaining the licensing differences between Google-Earth-standalone-install and Google-Earth-browser plugin: for a company that embraces simplicity on their homepage, a similar commitment to clear, attractive licensing to would-be business users would be most welcome.
p.p.s. As for the argument that it’s all the doing of the data providers (TeleAtlas, NAVTEQ, Digital Globe, et al ), if Apple can show recalcitrant music companies the light with iTunes, MS and Google surely can flex their muscle and convince their providers that there’s a win-win scenario here.
Totally agree. Plus they are loosing business.
We resell to small gps asset tracking dealers. When we tell them about the Google Maps and Virtual Earth licensing fee they about choke. They don’t have the volume to afford their stuff. So they don’t install it. Google and MS would more than triple their license income. If they made the switch you described.
James,
I’m about to demo the EarthBrowser Web Plugin at Adobe MAX for the RIA of the year awards next week. It might be just what you need for your mashup, it’s got a Javascript API with KML and Shapefile support and is very easy to code. It’s still in beta but I am hoping to release it in December. I haven’t got a prices fixed yet, but an internal app will be in the low hundreds rather than $10K.
Here is the link to a Live Demo which was completely written in Javascript in a few days.
Email me if you (or anyone reading this) would like to have an early API test key (mgiger at earthbrowser.com) to give it a spin on their own site.
I am far from an expert on this subject but have done some research recently. I understanding is that if you are publishing a Google Maps based app publicly it is free to use the Maps API but costs if implemented internally. VE API costs no matter what.
I was sent the following regarding VE licensing. Looking at the costs below it seems you should be able to publish a simple, non-driving direction enabled application for much less than $10K. FULL DISCLOSURE: The information was sent to me on 8/13/2008, so please check with Microsoft for more current information.
Begin email —>
In order to use the service in your agency, two parts are required:
Both of these components can easily be purchased off of your agency’s select or enterprise agreement via Microsoft Volume Licensing, which also will provide your agency with the best pricing possible. For budgeting purposes, I am authorized to provide you with estimated retail pricing (ERP) based off of today’s GSA schedule. Due to a licensing restriction from the imagery provider, Birds eye imagery is not available to US Government customers.
1) Virtual Earth Platform Access
· If you want to provide driving directions to points of interest via your agency’s website, I recommend the Professional Access Package. The pricing for this service is prorated monthly, and can be purchased through your Select or Enterprise Agreement. The ERP is $983 a month or $11,796 annually.
· If driving directions are not part of your solution, the Standard Access Package is available at ERP $316 a month or $3,792 annually.
2) Virtual Earth Platform Transactions
If the application using the Service will be supporting a public-facing internet site, accessible by unknown users, transactions are available in buckets 100,000 or 1,000,000 per month depending on your program’s requirements.
a. 100,000 monthly transactions are available for ERP $267 a month or $3,204 annually (1,200,000 total annual trx.)
b. 1,000,000 monthly transactions are available for ERP $2,106 a month or $25,272 annually (12,000,000 total annual trx.)
———— or ————–
c. If the application using the Virtual Earth Platform will support a Line of Business internal application with a known set of users (i.e., a call center, dispatch, etc…) then you can opt for the convenience of not counting transactions and procure the “per user” option of $7.00 a month per user.
This flexibility allows customers to purchase any necessary combination (a., b., or c.) of transactions to support the program mission depending on if the application is internal, external, or a combination of the two
pay for what you use via AWS like metered service – You can do exactly that with ESRI ArcWeb Services, plus you get nice flex and javascript apis.
I always ran into problems with people not feeling comfortable with a metered service. How does an agency with a fixed budget plan for variable cost, and what if they run out/over when they really need it.
ESRI is closing ArcWeb.
One thing I don’t agree with is your assessment that you have to pay before you write any of the code. You could easily develop your application on the Maps API (from Google) and then once you show the value, have “sold” it to your client, pay the licensing fees. Why is it the world thinks that everything should be free these days? These applications cost real money to run, and as you noted, both Google and MSFT are hamstrung by their license agreements with data providers. Sources tell me that both vendors have contractual minimums they have to pay for “in business use”. Maybe we should be complaining to teleatlas and navteq, why don’t they have “pay as you go” licenses to their data?
Bill: No one is going to hire me to develop an application and not pay for the enterprise license. You have to factor in the cost of the license before hiring anyone to develop an internal web mapping application. I can give clients tons of examples of “value” without them paying me to develop a website for them.
I don’t see anyone here saying this needs to be free. Where do you get that idea from?
Microsoft and Google are pushing their mapping APIs as developer tools, but in fact they aren’t. If they want them to be accepted as such, they need to rework the licensing to better address realities out there. This type of licensing worked in 1998, but in 2008 there is no way you can expect anyone to pay $10k to run an internal Google Maps application.
Everyone wants to push the blame on TeleAtlas or Digital Globe and that might have been correct in 2006. Today Microsoft and Google market their APIs as solutions and unless it is publicly available or the user has lots of money, it just isn’t feasible to release an internal Google or Microsoft web mapping application.
As usual I find myself in complete agreement with you James (and Brian T.).
Co-workers always say, “Why can’t you do something like that for us?” and I usually have to say because we can’t afford it. I’d love to be able to pay for what I’m using, rather than what I could you (assuming our company hired a ton more people and had a ton more clients).
Also, I find myself also agreeing with the GeoMonkey quite a bit. Might be time for some GeoMonkey shirts for everyone.
Dumb remark removed by dumb contributor! If I only had the brains of a GeoMonkey, sigh!
what do you compare? aren’t s3 and ec2 different concepts than providing rastermaps and api from goo and mic? how do you use s3 and ec2? can you tell it here in the blog?
would be hard for mic and goo to develop use-cases for their maps and apis because they don’t know what a geomonkey can do with it. they are not yet in that professional sector and “gis” in their case is more an attachement for connecting geospatial information to the user-profiles they have. storage and cloud computing are known concepts for amazon because they know what will use their hardware. a geoprocessing model may exponentially slow down your cloud. the traffic from server to client would be small.
James, right – and you have to ask why ESRI is closing ArcWeb Services – it seems to have everything you want, and yet people never flocked to the pay per click map service concept – I am suggesting because variable cost for using your web mapping app is hard for most people to budget for. It costs a lot of money to set up a pay as you go infrastfucture account management – it sounds good, but it looks like few people would actually use it. Would it be different for google or MS?
Also – while ArcWeb Services is closing, the concept may just be shifing to ArcGIS Online. For example, ArcGIS Online is free for internal use, but for external commercial apps you pay, and for anonymous user access, such as a public facing web app you sell to your customer, you pay ESRI based on transactions – e.g. the pay for what you consume model you are looking for.
Wow, what a great topic. Count me as one who would love better pricing. We use both Google and Microsoft internally without paying a license. We can’t afford the enterprise licenses and if Google wants to look the other way, who am I not to take advantage.
Don’t get me wrong, I’d be happy to pay, but we can’t justify the costs.
In theory VE uses a per transaction cost scheme while Google uses a per page view cost scheme. (beware since the transactions clip by at a tremenduous rate in a pay per tile plan) The problem is they both have a very high entry fee, at least for most small busines apps. Unfortunately they also both have a wonderfully complete base map which is hard to come by in a world wide application, so the client expectation is VE or GE without cost and the disappointment at reverting to public domain DCW, DRG, DOQQ, NAIP in a custom development scenario is hard to overcome. This, I assume, is why OpenStreetMap actually has a future as the “base map” for the rest of us.
And Brian is right, if I were the distant second play in the map api game, I would surely look at the free use/ad revenue model to attract developers to the dark side. Is there some problem with ad revenue on map searches that I am missing? Maybe it just isn’t lucrative enough to generate required cash flow to support the massive base map investment. Uh Oh – Is it possible that location just isn’t important enough to consumer choice to warrent adword type business models?
Being in the imagery bis for 12 years Watching all of this is interesting, I predicited that both VE and Google are going to have to come up with ways to help off set costs for highly expense base data. Imagery is expensive to produce and maintian and the only way to make this sustainable is through collaboration of public and private entities. The pay as you go which I monitored for a company I wont mention can get more costly than a lump sum cost and as one person said its hard for govenment to budget and pay that way. I watched one agencey do the pay as you go and once everyone knew it was there all these people started downloading lots of data and their bill was 27,000 in no time. Sustianablility is a hole different perspective, imagery is obsolete in a 1-2 year period now. Both GE and VE are trying to figure out ways to make this cost effective as right now its not. the wow factor is over.
So, how long before companies realize that it’s much cheaper to use openstreetmap.org and spend (part of) that money on field work?
To me, it’s the difference of selling a particular product instead of being a project driven company that make the licensing problematic. Companies like MapDotNet 1) sell one thing and 2) have an exclusive deal with Microsoft, so it works within this model. Who knows what a project driven firm might be doing next year or what a client wants until they engage them? Neither can afford a lot of time and money to speculate. It’s a lot of coin for small business to cough up just to buy in. That is where this license arrangement starts to break up.
I agree with a previous post, the non GIS people/clients are surprised to hear that the map they use for free is so expensive. That doesn’t change their expectations of the mapping technology or performance though.
Its been said already twice but I can’t echo it enough.
OpenStreetMap..
Use that with OpenLayers. The API is simple, easy and free.
If you are worried about security or speed, it is routine to setup your own server.
..or spend your $100K on a product where you can’t control the quality or performance…
OSM works for some clients but not all. Some really want the Google API because their users are comfortable with it. They might geocode against teleatlas and want the similar road network or tons of other reasons.
OSM comes off very “hacker” to many of my clients and they can’t get beyond that. I hope it doesn’t fall into what OSGeo is becoming.
James: I agree that OSM does not (yet) cover the necessities of all people.
However, case in hand: small enterprise that doesn’t want to (or can’t) spend $5K-$10K on map licenses for a pushpin mash-up. OSM is a no-brainer.
Let’s say your company developers are more confortable with the GMaps API. In this case, it’s the company’s call to decide whether to spend money on licensing, or on re-training their dev team on using OpenLayers. Or use mapstraction and load OSM tiles. In any case, the company should do a cost-benefit study on the alternatives.
On the “hacker” aspect of OSM: wanna go professional? Just talk to CloudMade (and/or GeoFabrik and/or AND and/or ITO). They will offer you professional support and services (on top of OSM data) for much less money than Google or Microsoft.
Do you really want something comfortable and non-hackerish? Spend those $10K you were talking about in the post.
In short: it is my opinion that lobbying Google and Microsoft to lower the prices of their licenses is not the way to go. Do you really want them to lower their license prices? Then stop bitching and switch to another map data provider. Supply and demand will do the rest.
The point of all this is that Google and Microsoft need to adjust their license costs to reflect the realities of the marketplace. Their tools aren’t being used they way the license assumes.
I use OSM and OL just fine so it, but clients of mine aren’t comfortable with OSM even if there was a “professional” such as CloudMade in the middle. I offer professional support for OSM without having to force them to pay another company. No matter how “professional” I am, some companies have serious problems with how OSM is run. It isn’t my job to “educate” them on OSM, that is OSM’s evangelists job. Same with Google and Microsoft, if they want companies to pony up the money for their licenses, they better start doing a better job at selling it than their current “enterprise” websites.
Bitching? Where do you see that? I see a clear case laid out for Google or Microsoft to support. We do use OSM where possible these days, but that doesn’t mean I won’t be giving feedback to Google or Microsoft where I think they are wrong. That statement of yours is the kind of stuff that turns people off of OSM and what I’ve been trying to fight through. Insulting companies that want to use Google or Microsoft (even if those reasons aren’t valid) isn’t going to gain any traction.
forgive me if I have misunderstood the licensing agreement re google earth .. but if you use a server inside a clients intranet to generate KML and the client then uses Google Earth Pro to overlay that data then there is no poblem ..
eg if the client has a $400 licence of google earth they are free to share that information
“”"” Share your Google Earth views and data representations with your clients as a KML, Google Earth‘s original file format. With your upgraded Pro subscription, you get additional measurement tools (square feet, mile, acreage, radius and so on), so simply select the points on the screen using your mouse and let Google Earth calculate the rest. “”"
I would like to buy 1 or 2 chicken wings instead of a dozen. Why can’t i have this?
The Google/VE APIs are nice but you can do similar things with free tools like OpenLayers etc, so nobody’s really paying $10K licence fees for the Javascript APIs, are they (unless they’re buying ESRI of course…)?
As for the data, does/could Google or Microsoft offer a geographically restricted contract, for example? After all, a real estate business with local maps of properties here in Wales hardly needs high-res satellite imagery of Wisconsin, does it? And it should be pretty easy for Google to restrict data requests to a given bounding box based on the app key.
Which I guess brings us back to James’s point about being able to pay for what you need to use – by volume or by area – rather than the whole thing.
Two thoughts:
a. OSM vs GE and VE comes down to the breadth, beauty, and convenience of base layers, the global model, which is also the major capital cost. The viewing software isn’t worth $10k – browsers are free, but the content costs money to build and serve. Google stumbling onto ad revenue is quite an incredible gift to the internet. I just wish it was applied to the map side of life and can’t really see what Microsoft has to lose by abandoning a losing revenue model.
I also don’t really see that OSM or World Wind can really keep up with the rapid expansion of the global models available to Microsoft and Google, especially as we see geography expanding rapidly from aerial views to street views, and now into the interior landscapes of facilities management and virtual life models.
b. kml alone doesn’t give the hooks necessary for interactive webapps so kml to Google Earth Pro is limited compared to Google Earth API. Google Earth is standalone and provides only primitive methods to interact with the model. What an app builder really wants is some way to wrap a user specific GUI around the model and interact with it. You really want to have a full range of event listeners attachable to any point, line, or polygon. If you just want to share some info layers GE Pro is fine, but as soon as you want the to interact with the model you have to use an API. Again, though, the API cost is nill compared to the global model.
Wonder if there’s a mechanism/market for re-sellers? Big company pays for the full monty from Google, then sells metered access to “their” Google data as a service to small companies, who build their apps on that basis. Not sure how you’d implement this in practice, but where there’s a market, there’s usually a way…
Overall very good points – I obviously agree with the “use OSM and OAM” points. One major need here is an OSM mapnik style that mimics Google’s Style. That would be an interesting comparison.
However, I also see some confusion between the “API” and the “base layers” – you don’t have to use OpenLayers to view alternative base layers, as @ivansanchez points out. The GoogleMaps API can load in other tile layers, and tools like Mapstraction can even make this easier.
So users can still have the “GoogleMaps” interface with alternative tiles.
“So users can still have the “GoogleMaps” interface with alternative tiles.”
Yes, but you will still have to have the $10k entry fee to use your alternative tiles with the Google base+API.
OSM mimic of Google style still doesn’t solve the problem of replicating the monstrous global model out there in the gazillion cpu Google Data Centers.
That monster global model really has only a couple of players – Google and Microsoft VE
MapQuest can’t keep up and of course USDOD NGA isn’t ever going to have a public API
Interesting that these posts seem dependent on global coverage. Perhaps the issue is that the GeoWeb hasn’t found it optimal adaptive resolution for delivering what most people are looking for? Jim_level18
@binarystar
The $400 license of GE Pro seems very reasonable, but that is just one license for a year. AFAIK, if you want to several users to use GE Pro you will quickly spend the $10K on that over time. I have heard that enterprise is at least $250K.
Eric
TheWebService offers pay-as-you-go mapping: http://www.thewebservice.com/webservices/
click through for code snippets, etc. It costs a few cents per map, but the only drawback is that more credits are consumed when you move the map, zoom in or out, etc., so if you’re not careful the cost could be a bit more than think. As far as I’m aware this is a limitation of the data licensing… but still well below the monstrous figures being tossed about here!!
Great post – although we’ve been working from the other side to get per user costs for tile services within the enterprise. Problem we run into is large enterprises need to be able to work within budgets and the transaction pricing for G and M does not work well. Since it is variable cost they do not know what the expense will be till the end of the year. Ironic that the up front cost for a minimum bundle scares off the SMB market as well.
I see this as an opportunity for the GIS and cartography professionals to further demonstrate their value. Leveragins open source data like OSM, OAM, GeoCommons, GeoNames etc. there is plenty of information to create your own base map and tiles. You can leverage Mapnik if needed or other design packages. We complain about the one size fits all cartography of the G, Y, and M’s of the world and this is the perfect opportunity to create custom base maps to really meet the customers requirements.
This is the way we did it back in the day and now with the abundance of open source data and tools you can do this at web scale. Further the work is reusable. For $8-10k I bet you could easily have a custom set of tiles you could reuse and tweak for many projects.
The problem is that the open source data sources are not backed up by the availability and monstrous bandwidth of Google and MS. You want aerials? Go to the county, state, or USGS and you can get aerials equal to or better than what MS and Google are carrying, for free or cheap, and with leaf off conditions. If you live in the right areas (like Maryland, Missouri, and a handful of other states), you can get these as a web service. But you will not get Google and MS throughput on those services. I see the cloud coming to ArcGIS server pretty soon.
Brett: “I see the cloud coming to ArcGIS server pretty soon”
Yes, every cloud has a server lining, after all.
@Brett definitely agree when we are talking about public facing applications there is an advantage to G,Y,M’s infrastructure. In the case of enterprise deployments the user base is usually considerably smaller than the whole Web, so having to pay $10k to amortize all that infrastructure often times does not make sense. You can use the open source web services or host the tiles yourself. We do that with blue marble on GeoCommons. With tools like tilecache and OL or ModestMaps it is not hard to do, more reliable because it is local, and cheaper. You can also leverage OAM (open aerial map) to stitch all that free imagery together if it has not been already done through the service. Then it is available through a single web service or again you can host it locally.
Interesting reading, our local ESRI GIS distributor seem to have changed their model for charging licenses as a result of the increased options as per the above. For the past few years our user group has been grumbling about the licensing charges here in NZ compared to the USA, and google/microsoft et al are becoming viable alternatives with the work Data Sciences are doing with data provision worldwide. This year things have changed with the recession and increased global service competition, just my 2cents from down under.
Cheers Duane Wilkins