Will Google Fonts ever go away?

Jeremiah Shoaf
Jeremiah Shoaf Posts: 11
in Type Business
I was reading through the mega-thread "Google Fonts: Your Questions, Answered" and this question was sort of addressed at the end but not really. I thought it could use its own discussion...

Google is notorious for shutting down services that no longer meet its business objectives. See the Google Graveyard.

My understanding is that Google Fonts was created to help improve search by having more sites use HTML text rather than image graphics. What happens when this goal is achieved? (I rarely see text as graphics on the web anymore).

It seems like the cost of serving web fonts to millions of websites has to extremely high, even for a company like Google.

What financial incentive do they have to continue the service? What are the chances of Google ever shutting it down?

I would love to hear the type community's thoughts.
«1

Comments

  • James Puckett
    James Puckett Posts: 1,994
    What financial incentive do they have to continue the service?
    Gathering data about font downloads. There has to be a use for that information.

    But even if GWF shut down it wouldn’t be a huge loss to the users. They could host the fonts themselves or use that other popular font hosting service.
  • Jackson Showalter-Cavanaugh
    Jackson Showalter-Cavanaugh Posts: 539
    As long as people use these fonts, google can continue collecting their data.
  • Jeremiah Shoaf
    Jeremiah Shoaf Posts: 11
    I never really understood how exactly the font usage data is that valuable to them. Knowing how many sites are using Open Sans is nice to know, but how can that data justify maintaining a huge server infrastructure to host fonts for millions of sites?

    Google Reader was shut down and I'd bet the data about what blogs people were reading was probably even more valuable than what fonts sites are using.

    It just makes me wonder what their motivation is for keeping Google Fonts running over the long term...
  • Cory Maylett
    Cory Maylett Posts: 248
    I have no way of knowing how they use it, but between Google Analytics and Google-hosted fonts, a whole lot of very specific site usage information is undoubtably being collected and analyzed by Google for possible use. Just off the top of my head, I can imagine them using that data to better target Google Ads and search results, among other things.

    Google doesn't think like other companies. A typical company creates products to sell. Google seems to think more along the lines of creating large user bases for their free servies, then using those user bases and the data obtained from them to support their more profitable ventures.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    edited September 2014
    (The opinions expressed are my own personal opinions, not those of any companies I consult for.)
    It seems like the cost of serving web fonts to millions of websites has to extremely high, even for a company like Google.
    Costs are relative.

    If I offer you something for $1 that you can get nearby for $0, its cost is extremely high. If I offer you something for $1M that you can't get anywhere else for less than $10M, the cost is extremely low.

    Whatever you think the costs of serving font files is, compare it to what you imagine the cost of serving images in Google Maps and the cost of serving video files in YouTube is. Does the costs of serving font files now seem high or low?

    I have no idea what any of the costs are, but I expect the cost of serving font files is relatively low.
    What financial incentive do they have to continue the service?
    The more people use the service, the more financial incentive Google has to continue it.

    Google makes some revenue from the Google Apps suite of productivity web applications. In serving fonts to millions of websites, fonts in Google Fonts are widely cached (only the first visit to the first site that uses a font suffers loading latency; all subsequent sites load the font instantly from the browser's cache.) The more the fonts are served to other websites, the better the font experience in Google's officeware suite, and the better the experience in all other Google products that use the fonts (such as Roboto, used in Google Maps and Google Plus) -- and equally the better the experience in all other websites that use the fonts.

    However, as anyone can read from the SEC filings, Google makes most of its revenue from web advertising. In serving fonts to millions of websites, Google Fonts makes the web better because web pages are more visually pleasing. The more the fonts are used, the better the web is, and then (a) the more people search the web, and the more people will click on Google's AdWords ads; and (b) the more peope read more webpages, and the more people will click on Google's AdSense ads carried on those pages.
    What are the chances of Google ever shutting it down?
    "On a long enough timeline, the survival rate for everyone drops to zero."

    Having said that, I personally see the web as still in its early stages of development; most human beings don't yet have web access. But they will. And having a collection of libre fonts for their language is important to their web experience. So I expect the Google Fonts project to continue for some time.
    Gathering data about font downloads. There has to be a use for that information.
    I don't think there is as much use for that information as you're implying.

    Since the 'Questions Answered' thread, I helped the Google Fonts product manager to write this privacy entry on the FAQ: https://developers.google.com/fonts/faq#Privacy

    I'm close to the project, and the uses are trivial and mundane. Afterall, there isn't very much information to begin with. The most interesting thing about knowing how often fonts are seen is to help to decide which fonts are worth investing more development in. Well, gee, what a suprise :)
    As long as people use these fonts, google can continue collecting their data.
    What do you mean, 'their' data?
    I never really understood how exactly the font usage data is that valuable to them. Knowing how many sites are using Open Sans is nice to know, but how can that data justify maintaining a huge server infrastructure to host fonts for millions of sites?
    I hope the above is sufficient.
    Google Reader was shut down and I'd bet the data about what blogs people were reading was probably even more valuable than what fonts sites are using.
    It can't have been that valuable, since they shut it down ;)
    It just makes me wonder what their motivation is for keeping Google Fonts running over the long term...
    I hope the above is sufficient.
    I have no way of knowing how they use it, but between Google Analytics and Google-hosted fonts,
    But Google Analytics isn't related to Google Fonts in any reasonable way, because the Google Fonts data is impersonal and scant.
    a whole lot of very specific site usage information is undoubtably being collected and analyzed by Google for possible use. Just off the top of my head, I can imagine them using that data to better target Google Ads and search results, among other things.
    I don't see how the choice of font could be relevant to either of those things.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    edited September 2014
    Google doesn't think like other companies.
    Google's founders don't think like other people because they went to Montessori schools...
    A typical company creates products to sell. Google seems to think more along the lines of creating large user bases for their free services,
    ...but that business thinking is very common: Many companies create advertising products to sell to advertisers, by creating audiences. If you charge a fee to the audience for the advertisig product, you limit the size of audience to those who can afford to pay the fee, and thus limit the revenue.

    This is the same business thinking behind free-to-reader newspapers that have been around for over a hundred years - http://en.wikipedia.org/wiki/Free_newspaper - and the same thinking that led to Twitter and Facebook and Medium and Reddit a billion other websites that carry adverts, take your pick of https://www.google.ch/search?q=cute+animals .

    Sometimes limiting the size of audience to those who can afford to pay the fee makes a more valuable audience. I don't see as many Rolex ads in the London Metro as in the UK Telegraph.

    But the Metro is more profitable.

    So, giving away advertising products is hardly new, and it always involves a tradeoff. The web doesn't eliminate that equation, but it changes it because the potential audience for zero price advertising products is so huge.
    then using those user bases and the data obtained from them to support their more profitable ventures.
    User profiles are not related to Google Fonts; they require at least a Google cookie, and likely signing in to a Google Account.
  • SiDaniels
    SiDaniels Posts: 277
    If the service were ever shuttered, I'm sure Google would give customers a few weeks to switch to a paid service or self host. The fact that all of the fonts are open source and self-hostable likely provides adequate insurance. People will grumble, but you get what you pay for.
  • Jeremiah Shoaf
    Jeremiah Shoaf Posts: 11
    Thanks for the thorough response, Dave. You definitely know a lot about this topic.

    1) I agree that the whole "data mining" argument probably doesn't hold up.

    2) You compare Google Fonts to YouTube and Google Maps but I don't think that is an apt comparison since both of those services are monetized through advertising. I see no real way for them to monetize Google Fonts.

    3) I understand the Google Apps caching argument but I would imagine people who use Google Apps are mostly return users so the fonts would be cached regardless. First time visitors might see a slight load time improvement I suppose. Still hard to imagine this being a good reason for them.

    4) The "making the web better for everyone" argument is interesting. It just seems like such a lofty and hard to measure goal. "Open Sans is so beautiful. Now I'm going to search the web more!".

    5) I don't doubt that having libre fonts available that support the languages of the world is important. I just don't see why Google needs to host those files for every website.

    6) Also it's not just the server costs. It's also the internal resources they devote to maintaining the project. Every resource they devote to this takes away resources from another potentially more profitable project.

    I guess overall I just feel that it's strange to see Google in the font hosting business. To me it has that feeling of one of those Google projects that engineers dream up with lofty goals and then management says "wait this doesn't fit into our core business objectives" and then axe the project.
  • Jeremiah Shoaf
    Jeremiah Shoaf Posts: 11
    SiDaniels:

    I agree. Also even if users didn't switch to another service or self host, the worst that would happen is the fonts would default to the next available font in the CSS font stack, most likely Arial or Georgia. I'm sure most non-designers wouldn't even notice.
  • Jack Jennings
    Jack Jennings Posts: 151
    For what it's worth, Dave is correct that serving fonts is not resource intensive for a business the size of google. There are many services out there that are freely hosting much larger assets like javascript libraries, HTML files, etc. Maintaining a large-scale content delivery network (CDN) is a generally solved problem and not something that's pulling top engineering talent away from their core business.
  • Jeremiah Shoaf
    Jeremiah Shoaf Posts: 11
    Maybe the server costs aren't that much for a company like Google. However, I'd bet the same could be said for other Google products that got shut down. I guess it's just a matter of what it costs them (which probably isn't much) vs what they get out of it (which doesn't sound like all that much either). Apparently if the service is still running they must be getting something from it to make it worth their while though.
  • Cory Maylett
    Cory Maylett Posts: 248
    ...a whole lot of very specific site usage information is undoubtably being collected and analyzed by Google for possible use. Just off the top of my head, I can imagine them using that data to better target Google Ads and search results, among other things.

    I don't see how the choice of font could be relevant to either of those things.
    It's not the choice of a font that's relevant. When a visitor to a website initially loads a page containing a Google font, a request is made to Google's servers to serve that font file. Like any server, detailed information on where those requests originate, their frequency, traffic patterns/cycles, etc, is stored. This information provides Google with a behind-the-scenes peek into that site's visitor usage patterns.

    One site at a time, this data has little value, but the aggregated data from hundreds of thousands of sites using their fonts could be a different story. This might be especially true when that data is combined with similar information obtained from other Google services, like Google Analytics and Google searches. When compared to those services, the relative value of the data obtained from their fonts would likely be limited, but given the relatively low cost of serving these files (along with their other potential benefits) I can easily imagine benefits outweighing costs.
    A typical company creates products to sell. Google seems to think more along the lines of creating large user bases for their free services,

    ...but that business thinking is very common: Many companies create advertising products to sell to advertisers, by creating audiences.
    The already existing models you listed mostly build audiences that they rent out to others. These aren't new business models, but their resemblance to what Google does is superficial and, even, primitive.

    Google seems more interested in the wealth of information that can be obtained when hundreds of millions of people use their products. Aggregating and automating the mining this enormous data set for predictive patterns with commercial potential on the scale that Google seems to be doing is new. They aren't the only ones headed down this road, but their information gathering and analysis capabilities are unequaled. It's definitely different from, like I wrote, a typical company.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    (Sorry, I wrote all this like 18 hours ago but somehow didn't post it :)

    Personal opinion only:
    The fact that all of the fonts are open source and self-hostable likely provides adequate insurance.
    I agree
    I see no real way for them to monetize Google Fonts.
    I have explained how they already monetize the fonts.
    hard to imagine this being a good reason for them
    You can have a website without videos, but you can't have a website without text. And text means web fonts. So every Google product (indeed, every web site and web app) needs web fonts, and Google could choose not to provide the fonts to millions of other websites, but the experience of the fonts in Google's products would then be worse.

    And providing the fonts to millions of other websites is a 'complementary good' for Google's AdWords and AdSense products - http://en.wikipedia.org/wiki/Complementary_good ("more of each good will be demanded whomp"
    The "making the web better for everyone" argument is interesting. It just seems like such a lofty and hard to measure goal.
    Measuing the adoption of web fonts would require a copy of the web. Oh, wait. ;)

    This measuring was presented last year at ATypI by David Kuettel, the co-founder and lead engineer for GF, and John Giannopoulos, "Director of Strategic Alliances" at Monotype, and will be re-presented at ATypI again next week in Barcelona. I'll be there.

    If you are saying that its hard to measure the impact of typography, psychology scientists do attempt to measure it. SiDaniels' colleagues at Microsoft are pushing the state of the art in this scientific understanding forward every year, and Monotype has also done studies in this area.

    I suppose that is because MS is also interested in justifying the millions of dollars they spend on font and font rendering technologies with science; and Monotype uses the studies for sales and marketing.
    I just don't see why Google needs to host those files for every website.
    I just don't see why you don't accept the reasons I am giving you :)
    Also it's not just the server costs. It's also the internal resources they devote to maintaining the project. Every resource they devote to this takes away resources from another potentially more profitable project.
    What can I say? Some people just love fonts :)
    I guess overall I just feel that it's strange to see Google in the font hosting business. To me it has that feeling of one of those Google projects that engineers dream up with lofty goals and then management says "wait this doesn't fit into our core business objectives" and then axe the project.
    I have explained how the project is aligned with the core business objectives.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    Personal opinion only:
    When a visitor to a website initially loads a page containing a Google font, a request is made to Google's servers to serve that font file.
    Yes, the initial page load. After that the CSS is cached for a day, and the fonts for a year. Google sees one CSS request per browser per day.
    Like any server, detailed information on where those requests originate, their frequency, traffic patterns/cycles, etc, is stored. This information provides Google with a behind-the-scenes peek into that site's visitor usage patterns.
    One CSS request per browser per day is a very small peek.
    One site at a time, this data has little value, but the aggregated data from hundreds of thousands of sites using their fonts could be a different story. This might be especially true when that data is combined with similar information obtained from other Google services, like Google Analytics and Google searches.
    How can the data be combined in a meaningul way? The data you are suggesting it could be combined with is so much richer, so how is this meagre piece of font request metadata helpful?
    When compared to those services, the relative value of the data obtained from their fonts would likely be limited, but given the relatively low cost of serving these files (along with their other potential benefits) I can easily imagine benefits outweighing costs.
    The already existing models you listed mostly build audiences that they rent out to others. These aren't new business models, but their resemblance to what Google does is superficial and, even, primitive.
    That's what any advertiser does; direct advertising has been doing "list segmentation" for 100+ years. https://www.google.ch/search?q=list+segmentation
    Google seems more interested in the wealth of information that can be obtained when hundreds of millions of people use their products. Aggregating and automating the mining this enormous data set for predictive patterns with commercial potential on the scale that Google seems to be doing is new. They aren't the only ones headed down this road, but their information gathering and analysis capabilities are unequaled. It's definitely different from, like I wrote, a typical company.
    Maybe 15 years ago.... but today it is very typical for every company to do this sort of thing. http://www.forbes.com/data-driven-business/
  • Dave Crossland
    Dave Crossland Posts: 1,429
    Maintaining a large-scale content delivery network (CDN) is a generally solved problem
    I agree. Given how many CDN service companies are out there, I'm surprised that foundries who complain about their limited sales channel options don't set up their own web font services. If GF did go away, it would be easy to do the same.
  • Jack Jennings
    Jack Jennings Posts: 151
    Just because it's solved doesn't mean that it's trivial to set up. Especially for shops with small margins, I imagine that hiring a proficient software engineer that can build and deploy the software is quite out of reach.
  • Thomas Phinney
    Thomas Phinney Posts: 2,888
    I agree. It's not rocket science, but it is not trivial.
  • James Puckett
    James Puckett Posts: 1,994
    More importantly, it’s distracting. Because if you set up a web font service you at least one more server to secure,update,test,etc. And then you have to market it. And handle customer service for a 24/7 web service (good luck outsourcing that to an Indian call center).
  • Dave Crossland
    Dave Crossland Posts: 1,429
    You pay the CDN to do that maintenance, James & Jack
  • Deleted Account
    Deleted Account Posts: 739
    "...the fonts would default to the next available font in the CSS font stack, most likely Arial or Georgia. I'm sure most non-designers wouldn't even notice."

    Indeed. This rather seems to negate Dave's apparent arguement. If non designers cannot tell the difference between Google or default fonts, much less between a google font used by one advertiser vs another, as is mostly the case, prettier and faster don't matter.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    If non designers cannot tell the difference between Google or default fonts
    Non designers can tell the difference between Google or default fonts as much as they can between WebType.com fonts or default fonts.
  • Deleted Account
    Deleted Account Posts: 739
    Lol, but that's neither the comparison I'd think google's ad,s for a 33% market share, should want to make, nor one Webtype actually suffers from, by the looks of it, or the books on it. ;)

  • Dave Crossland
    Dave Crossland Posts: 1,429
    The books don't lie, so prettier and faster matter
  • Deleted Account
    Deleted Account Posts: 739
    ...and then faster dies and pretty wins.
  • Dave Crossland
    Dave Crossland Posts: 1,429
    LOL Frank

    Which way is that money going??
  • LeMo aka PatternMan aka Frank E Blokland
    LeMo aka PatternMan aka Frank E Blokland Posts: 723
    Eventually it went to the waitress, IIRC. ;-)
  • Jeremy Dooley
    Jeremy Dooley Posts: 67
    edited January 2015
    .
  • Fontfruits
    Fontfruits Posts: 51
    I thought this an interesting topic to revive...as google fonts has grown so much in almost 10 years. They certainly don't appear to be going away....and Frank certainly didn't get tarred and feathered  :#

Categories