My Love/Hate Affair With Typekit

GEORGIA and Verdana, Lucida and (to a lesser extent) Arial and Times New Roman have served us well. For fifteen years, these cross-platform default fonts have been faithful stewards of our desire to read, write, design, and publish web pages. Yet we designers have always wanted more. As far back as 1994, we hoped for the day when we could brand our layouts as magazine and poster designers do, by setting our pages in Franklin or Garamond, our headlines in Futura or Rosewood. And since 1998, CSS2 has provided a standard way to embed any typeface, not just the fab five, on a web page.

In August, 2007, CSS co-creator and Opera Software CTO Håkon Wium Lie wrote CSS At Ten, reminding us that CSS provided a mechanism by which actual font files could be linked to and retrieved from the web. Soon after the article was published, “web fonts” discussions started popping up at interactive design festivals and my friend Jeffrey Veen got the idea for a product that would get web fonts happening without running afoul of inconsistent browser support, multiple format hangups, or type designer licensing agreements and piracy concerns.

Speeding up design acceptance

While browser improvements and web standards alone provided multiple partial solutions, Typekit offered a complete solution that just worked. And the people behind Typekit (including Bryan Mason and Jason Santa Maria) did everything right: they reached out to the type design, graphic design, and standards-based web design communities; they worked with vendor after vendor to offer as many fonts as possible; they spoke everywhere, marketing their venture one lecture and even one designer at a time.

Typekit excited the web design community about type and proved that licensing and hosting web type was a viable business, providing options and convenience for designers and their clients, while bringing new revenue to type designers and protecting their intellectual property.

Typekit is the tipping point

Publicly and truly, I support Typekit because it is getting us to the world of web fonts faster. We could wait indefinitely for type vendors to agree to industry-standard licensing terms and font formats. We could wait far longer for IE, Firefox, Safari, Chrome, Opera, Opera Mini, Mobile Safari, and the rest to support the same font formats. (Currently Firefox supports WOFF and TrueType, Safari and Chrome support TrueType, MobileSafari supports SVG, IE supports EOT, and on, and on.)

But with Typekit, we don’t have to bother our pretty little heads worrying about these inconsistencies, and we don’t have to sit on the sidelines, waiting for all font makers and all browser makers to support a single standard format.

Platforms and performance

Typekit works, and that helps web designers and type designers take “web fonts” seriously. Typekit’s success is even helping to make web designers and type designers more aware of platform problems that can make fonts hideous on various platforms. Georgia was designed for the screen. Garamond was not. Moreover, platforms vary the way they hint fonts (Apple throws out hinting altogether, Microsoft over-hints) and the way they render them (from purely pixellated to at least three varieties of sub-pixel anti-aliasing), making a font’s appearance on a given user’s system hard to predict.

If not for Typekit, we might have had to wait years for most or all type designers to license web fonts. Only then would we have discovered that body text set in anything other than Georgia and Verdana pretty much blows on many Windows OS, browser, and monitor combinations.

Thanks to Typekit, we all know about the problem, and type designers are re-hinting their fonts, and in some cases redesigning them for the screen.

For all this I and all of us can be grateful to Typekit.

They also understand that designers will only use “web fonts” if they have access to the fonts they need. Just as a huge selection enabled iTunes to dominate online music, Typekit’s makers know their service must offer pretty much every good typeface out there—and they are working on it.

Renting versus “owning”

All this said in Typekit’s favor, I have mixed feelings about their product because I’d rather buy a web-licensed font than rent it—and Typekit’s success at establishing the viability of a rental model means that individual type foundries will also rent their fonts—and those who succeed at renting their fonts to web designers may not be inclined to sell.

Of course you never really own the fonts you buy—you simply license their use. So the analogy of owning versus renting doesn’t exactly hold true. But a one-time font purchase as a line item in a design budget is easier to explain and sell to a client than an ongoing rental charge.

Web Standards and @font-face

My other qualm has to do with a preference for pure web standards over product-assisted web standards. I don’t know if my preference is ideological or just the way my mind works (or fails to). But, given my druthers, I’d rather see millions of websites using standard @font-face to link to self-hosted web-licensed fonts than see that same number of fonts using a service—even a brilliant service created by friends for whom I wish continued, deserved, great success. It must be a quirk of mind; there’s no other logical explanation for this preference.

For those who share this bias, possess the properly licensed fonts, and don’t mind using FTP and writing a little code, the CSS @Font-Face Generator by Font Squirrel provides an exceptionally easy way to automatically generate the font formats necessary to take all browsers (including mobile) into account—complete with automated Cufón backup and your choice of best-practice @font-face code strings.

See also FontSpring.

Read more

40 thoughts on “My Love/Hate Affair With Typekit

  1. I conducted a small experiment comparing Cufon, @font-face, and Typekit. I have to say Cufon did the best job in rendering across platforms. Both Typekit and @font-face have some issues on IE that makes me shy away from using them on projects that have an audience beyond the Mac crowd. As it stands right now, though, I still like Typekit over @font-face, especially since I’ve already paid for it :P

  2. I agree with most of the points presented above, but I feel like it’s worth noting that from a performance perspective, Typekit is actually an advantage as more people continue using it.

    Someone correct me if I’m wrong, but it’s much like loading jQuery from Google’s CDN. Font files can be quite large, so what’s wrong with having one central resource serve them up so they’re cached across page/site loads? If I can avoid an initial 80~110kb bump on a first time page load, I’d gladly take it.

  3. Jeffrey — I’m right there with you. I love what Typekit is doing — especially if it truly leads to improvement of the many platform-specific rendering issues — but there’s also a nagging sense that there should be a more direct method for legally and logistically managing font usage on the web.

    I’ll immediately contradict myself, though, by saying that in spite of my knee-jerk, negative reaction to the subscription model, I’m actually not totally sure that’s the wrong thing. I certainly don’t like the chain of dependencies that get created (it’s hard enough to maintain the integrity of a site’s design over time!), but I am also eager to see how the livelihoods of our industry’s many talented type designers can be best supported while making their wares available for widespread use. (And that’s wares, not _warez_ ;) )

    It seems like a bit of a puzzle, and I’m really glad there are folks out there pushing towards a solution. So cheers to Typekit.

  4. Great points, and I agree mainly with the difference between a one time purchase vs ongoing fees. It ends up as another vendor to keep track off, pay,and remember your password for. That said, Typekit does have a impressive font collection and a rock solid UI to boot.

    I myself have used @font-face as it’s free, but @David’s experiment has me rethinking Cufon by way of CSS-Squirrel’s Generator (unwanted Cufon generated markup aside). Excellent resource.

  5. I am just getting into typekit myself. I like it so far, but I have that nagging feeling in my mind that there is a feeling of giving away control of a crucial part of your site to a service you are locked in to and don’t have any direct control over.

    On the plus side though. Just another thing not to worry about. That is life really, deciding what you want direct control over compared to how much hassle you want to deal with doing it yourself.

  6. Wow, I cant believe you posted this today, but it doesn’t matter. The important thing is that the word gets out.

    I totally agree with you about using standards to get more type faces on sites. However, I would recommend using Cufón. Dave Shea uses it to draw type using SVG rendering the font of his site inaccessible.

    This is why I suggest Typekit and the upcoming Font Deck. They implement accessible forms of font.

    Until, the new font working group can get browser makers to move in the same direction I don’t know that I would suggest doing font embedding yourself.

    Yes, its time to use new type faces, but we need to be careful if we implement them ourselves. Services like Typekit will keep our sites from breaking if platforms change.

    Standards are the future but unfortunately we need to use services today.

  7. In my experience, I simply mentioned where and how to buy the font, and the client understood it as just another fee to pay (like hosting), no problem.
    Though from a “client usability” point of view, I agree with you, I’d rather pay for the font myself once and include it as a line item in an invoice.
    Maybe it depends on how much you, as an agency, are used to charging a “monthly fee” to clients.

  8. Thanks, Jeffrey. I really appreciate the kind words. You are right in noting that our goal at Typekit has been primarily to bring fonts to the web, and to do so quickly. We felt the time was right, even if the technology and business models were – and are – still emerging. Iterating in public view can be tricky, but I’m glad we’re doing it this way, to be honest.

    Your music metaphor is a good one; more than a few people have called us the “iTunes of fonts.” Our original model is actually more like Spotify – pay a subscription fee and get access to a full library of content. But it’s also worth pointing out that we’ve embraced the iTunes-style “buy a one-time license” model as well. For example, you can buy webfont licenses at FontShop, and bring that license to Typekit to use instantly with browsers that FontShop doesn’t support. More details on that here on our blog. We’ll be announcing a bunch more foundries using this method of licensing soon.

    To your last point, I don’t really think Typekit is an example of “product design over web standards.” We use 100% standards compliant markup, style, and script. Go ahead and view source on the code that’s served to websites using our service; it’s all there. Instead, think of what we’re building as a service that sits along side web standards, helping designers and developers focus on creative solutions rather than workarounds and hacks, no matter how bullet-proof they are. It’s like how so many people use jQuery hosted on Google’s servers now. Everyone who includes the link in their page automatically has the latest stable version, properly minified and gzipped, served from data centers around the world.

    That’s what we’re offering as well – you can use our platform as a subscription service with a library of fonts, with licenses that you bring to Typekit, or both.

    Anyway, thanks as ever for raising the bar and hosting the debate. I always enjoy how you think about things.

  9. To your last point, I don’t really think Typekit is an example of “product design over web standards.” We use 100% standards compliant markup, style, and script.


    You’re absolutely right of course, and I would have done better to phrase that as something like “direct versus assisted web standards.” I’ve changed the wording accordingly.

    The “buy a one-time license” information is exciting, too—thanks for sharing that here.

  10. Very nice write-up, Z. I have an unhealthy obsession with type and am glad to see more options and discourse about it springing up on the web.

    I want to just comment quickly about the use of Cufón. It’s a great alternative to using different typefaces on your pages — with a few caveats. Many foundry license agreements don’t allow for this kind of embedding (even fewer than permit sIFR). And it’s important to understand that Cufón doesn’t produce actual font files, it grabs the outlines of the original font to render the characters on a page. This throws out a large amount of work that type designers do to create fonts (hinting, kerning, etc.), and won’t allow a design to take advantage of font-related features as they become available in browsers (like OpenType features). Because Cufón isn’t a font it bypasses the built in type rendering that browsers have. Cufón is useful for certain things and is a really interesting option, just make sure you understand it and are using it legally with your licenses. :)

  11. Typekit and web font licensing aren’t controversial when speaking of personal blogs or sites that don’t generate a lot of traffic. Typekit’s entry level pricing is affordable, and can make a big difference in the visual appearance of a web site.

    However, where the wheels start to fall off is with sites of scale. With Typekit, you would not only be entrusting over your site typography with a third party, but you’d also be paying a considerable amount more. How expensive that can get is murky, for Typekit suggests you contact them for “enterprise” service pricing.

    The web font licensing of Font Shop is also troublesome. Try purchasing any of their web fonts and select the average amount of monthly traffic your site receives (or you expect to receive). The cost for a site of scale is well into the thousands of dollars.

    Now, I don’t know about everyone else, but I don’t want site popularity to dictate site design. Rich typography shouldn’t be confined to low-to-medium traffic web sites. The whole web should benefit from it. And, when you think about it, it’s actually the top-tier web destinations (and their audiences) that stand to benefit most from improved typography.

  12. My newfound bias notwithstanding, and having used web type with CSS @font-face on a number of projects (both do-it-myself and via Typekit), I find Typekit a pleasure and a relief.

    Jeffrey, If I may use an experience we share as an example: Blogging became a little more about the writing when I switched to WordPress from hand-editing HTML pages (like you did). With Typekit, designing is a little more about the type.

  13. My concern with having a third party serve the fonts is the lack of ultimate control. This may just be an issue I have that warrants a visit to a therapist rather than Jeffrey’s blog, but the… doubt? worry? persists.

    I think TypeKit is performing a great service. My issue isn’t with them. It’s with the foundry that, for some reason totally unrelated to anything we’ve done or will do, decides to pull its font–the very one we happen to be using, say–from TypeKit’s catalogue. We’ve seen this time and again, whether it’s wrangling between news organisations and Google, music companies and, well, everyone on the planet, and most recently Amazon & its strong-arm tactics with some of its suppliers. It happens all the time and will happen again. Given the unbelievably tentative entry of foundries into the world of web fonts, I can virtually guarantee that it’ll happen here, too.

    The issue for me, then, isn’t a lack of trust in TypeKit–after all, they & I actively maintain a partnership when I decide to use them. It’s an issue with a ‘chain of trust’–that Foundry X isn’t going to break off its relation with TypeKit & thus screw up my designs in the process.

  14. I currently use Typekit on my own blog, and I have to say I love it. However, I am with you on the preference for @fontface. I would rather use this method, but until the issues with licensing and cross-browser support for different font types (EOT etc) are solved, I think Typekit is a good stopgap. Relying on javascript solutions for fonts isn’t a nice way to do it, but it’s a great way to progressively enhance. As long as you have reasonable fallbacks in your fontstack, I don’t see a problem.

  15. But isn’t it true that Typekit is not supported in both FF3.0-5 and Chrome 3.x? My colleagues and I have taken to using Cufón simply because of its cross-browser functionality. I really want to use/like Typekit, but if two relatively modern browsers don’t support it, that makes it a non-option for a lot of us…lest we have clients coming back using random browsers saying things don’t look right. (Please correct me if I’m wrong about Typekit support in those browsers!)

    That’s the frustrating thing about these various approaches. None of them seem to cover you functionally and legally.

  16. My own thoughts echo yours. I love the ability to have more choices in type but have not been able to get past the whole “renting” aspect. It’s also been a point that has come up with several clients who’ve opted to not pursue the use Typekit because of the monthly fee.

    It’s good to hear that there are some changes coming soon that will address these concerns.

  17. The whole web should benefit from it. And, when you think about it, it’s actually the top-tier web destinations (and their audiences) that stand to benefit most from improved typography.

    Agreed. Which is why a tiered fee is the fairest way to associate costs. We used three broad levels for FontShop licensing because we didn’t want a personal blog to pay the same as CNN. Pageviews is the simplest figure for distinguishing top-tier web destinations from smaller sites.

  18. Typekit is really an awesome start. I see the concern about ‘renting’ the font, but most of us rent a server because the economics make sense. Given the correct ‘critical mass’, I think we’ll see the same thing happen with web fonts. I can also understand the issue of recurring yearly billing being seen as stopper, but I think there’s a model for that as well; after all, the client has to renew their domain every year (they don’t ‘own’ it any more than a Typekit customer owns the fonts), and the individual costs could all just be rolled into a single service fee.

  19. Ray, that’s a valid concern for sure – it’s one we had when we were creating the service. We have contracts with each foundry we’ve partnered with to ensure the availability of their fonts. That is, a foundry could leave Typekit for any reason, and we’d stop offering their fonts to our customers, but customers who are using those fonts will always have access to them. The ‘chain of trust’ extends to our foundry relationships.

    Grant, just to be perfectly clear: Typekit uses @font-face to serve fonts. The only reason we use Javascript is to negotiate support between different browsers, as well as for a few interesting features we’re developing right now and will be launching soon.

    watermelonkid, Typekit supports any browser that supports @font-face, with the exception of Opera (and that has to do with bugs in that browser’s webfont support.) Have a look at the full browser support here.

    Todd Dominey, I’m sorry if our pricing seems murky. We’re trying to present our customers with a service that scales as their sites do. Sites that get millions of visitors understand that traffic comes with a cost, either implicitly or explicitly. Ours does, as well, even if it’s only the geographically distributed bandwidth we provide. Thankfully, as our service and our customers grow, the economies of scale mean we can provide more and more for less.

  20. I think the rent model allows fonts that would usually be quite expensive to be much cheaper. Typekit has nailed that model and as you said has been a leader in bringing real web fonts to the masses.

    Typekit and other services like it will be around for a long time to come, they are enablers for a much prettier web.

  21. Of course you never really own the fonts you buy—you simply license their use.

    Maybe it’s time for Open Source fonts to make an entrance; somehow if the design community could raid their collective piggy banks and pay a type foundry to license their font in a GPL fashion? Maybe a bit too pie in the sky…

  22. Hi Jeffrey Veen

    You said “Typekit supports any browser that supports @font-face, with the exception of Opera (and that has to do with bugs in that browser’s webfont support.) ”

    Now that Opera 10.50 for Windows is final, we’ve retested the test case you sent us by email (thanks for that) and believe that it’s working fine. Have you found more problems?

    Opera 10.50 Mac and Linux aren’t final yet, so we’d love to know if there are bugs that we can iron out.

  23. I hear what you’re saying on the subject of renting versus owning, but one of the advantages that web services such as Typekit and Fontdeck (it’s coming soon I promise) have is that readers always get the latest version of the font.

    Right now many typefaces are in a state of redesign, whether that’s just improving hinting or actually redesigning letterforms and adding new glyphs. It’s important that readers get the benefits of these improvements as soon as they are available. There has never really been the concept of updating a font (in the way that we’re used to doing with software) but that is one of the things that font services can offer, which you can’t really get otherwise.

  24. Great post, as always. I’m giving a presentation tonight on Web fonts and came to a similar conclusion. Typekit is a fantastic and brilliantly executed platform that delivers some of the best licensed fonts available. They’ve done a fantastic job of reaching out and crafting a solid site, system and support.

    The danger, to me, is what happens if Typekit (or Typotheque, with an analogous model) ever goes out of business? And what happens if clients don’t want to “rent” fonts? In fact, I’m not sure anyone knows how to charge clients for font “rental” quite yet. Is it like hosting? Or is it more like a nicety that designers throw in, akin to having a good copy of Photoshop.

    From a licensing perspective, however, I think Typekit is the way forward. Using FontSquirrel’s (excellent) @font-face generator, you can easily upload any fonts you possess (regardless of the EULA) and create a well-formed kit that supports all browsers, as you note. But, in the process, you expose the TTF to free download. That’s not an acceptable solution for foundries, designers, or individual type designers who seek to fund their work.

    One solution may be something like Fontspring, where you license fonts from $0.00 to $40.00 for a particular domain or unlimited domains. This ensures the integrity of the font files, provides some security for the fonts, and allows you to have full control over where they’re being served.

    For now, I’m liking Typekit, which supports standards and works exceedingly well. We’re just at the beginning of this. New models will continue to appear. And designers and developers will need to take baby steps with @font-face to ensure their content displays legibly as well as beautifully.

  25. Regarding your last point, I feel the same way. If a standard is implemented (@font-face), then why would we use a framework that accomplishes the same outcome? I understand the browser issues … but these are only temporary.

  26. I’m happy to see that one of my favorite type designers seems to be allowing people to make their own @font-face embeddable versions of his wonderful Larabie Fonts Collection.  His proposed workaround oozes good karma: “Change the font’s filename to something difficult […] to guess.”

  27. Bruce, you are right and I stand corrected. The latest version of Opera for Windows does, indeed, fix the bugs we had uncovered earlier. There are still a couple roadblocks for Typekit support, but we’re looking into alternatives and should have a solution soon. Any chance you guys will support WOFF? We’d love to start sending those files to Opera in addition to Firefox 3.6.

    Richard, great point! We update and upgrade our fonts continuously, and since we serve them centrally, our customers always have the latest versions. The selections from FontFont, for example, have been updated four times in the last few months with improved screen hinting.

  28. Currently Firefox supports WOFF and TrueType, Safari and Chrome support TrueType, MobileSafari supports SVG, IE supports EOT, and on, and on.

    This is misleading. Firefox, Safari, et al. support not just “TrueType” but also “OpenType” — meaning OpenType fonts with either TrueType (.ttf) or PostScript/CFF (.otf) outlines.

  29. @grant
    but until the issues with licensing and cross-browser support for different font types (EOT etc) are solved, I think Typekit is a good stopgap.
    The issues are not as great as you might imagine. Take a look at one of Font Squirrel’s packages and try their @font-face generator.
    It’s mostly a question of file conversion – which is an automatic process.

  30. Richard:

    Over the past two days, the Font Squirrel @Font-Face Generator has been freezing and crashing for me without ever generating files. This happens in Firefox, Chrome, and Safari in Mac OS X 10.6.2 on two different iMacs. A temporary problem? Something funky in my system (i.e. the typical non-reproducible Zeldman-only bug experience), or are other folks experiencing this sometimes, too?

  31. @Jeffrey – Well, that isn’t good. We’ve heard a few reports of crashes, but nothing specific we can use to diagnose. Does it just freeze? I’m a bit concerned that the processing power it takes to convert these thousands and thousands of fonts is taking a toll. Any specific info you can provide would be helpful. So far I have been unable to reproduce the problem.

  32. Thanks for once again articulating my thoughts for me Jeffrey. There was a need, typekit filled it, it worked but it charged! Enter Font Squirrel for those of us working on lower budget stuff.

  33. My problem with typekit is the pricing structure, the jump for being able to use it on a poxy 5 sites to being able to use it on a far too many 40 is insane, there needs to be a middle ground.

  34. Jeffrey Veen,

    glad that the Opera bugs are fixed!

    Regarding WOFF, we’re generally sympathetic to its aims. Which means (of course) that I can’t promise anything but we like it.

  35. I’m with you. I prefer @font-face to relying on JavaScript and pay licenses. I love the FontSquirrel generator, though I like to modify the CSS it produces to support multiple weights and styles using the same font name.

  36. Like most folks here, I agree with you; in fact, I was keeping my mouth shut, since most folks were quite in love with TypeKit.

    I honestly have no qualms paying extra license fees, but I do have qualms with relying on third parties services on the speed and agility of my servers… not to mention the proprietary reluctance for anything flash, especially in the iPhone age. That kills TypeKit and sIFR for me and my department, and with 13 companies worldwide, we have a decent reach. I’ll stick with Font Squirrel and @font-face for now until a solution as universal, simple and fundamental as background images for my css.

  37. @Jeffrey – I too have been experiencing some downtime on the Font Squirrel website. Seems to be working now though.

    As for my preferred font embedding solution it would have to be @font-face. It’s just so clean and simple.

  38. Awesome stuff
    Typekit is a fantastic and brilliantly executed platform that delivers some of the best licensed fonts available. They’ve done a fantastic job of reaching out and crafting a solid site, system and support.

Comments are closed.