23 Comments

Summary:

Without realizing it, you are a guinea pig for Google’s engineers. Gmail uses a slightly different shade of blue for its links — this one, #2A5DB0 — than the main Google search page, which uses #2200CC. Google’s engineers wondered if the difference in link color would […]

gmaillogoWithout realizing it, you are a guinea pig for Google’s engineers. Gmail uses a slightly different shade of blue for its links — this one, #2A5DB0 — than the main Google search page, which uses #2200CC. Google’s engineers wondered if the difference in link color would alter the rate of clickthroughs — is there an ideal shade of blue that encourages people to click links?

To find out, Gmail’s users were randomly tested with 40 different link colors, ranging from blue-with-greenish to blue-with-blue-ish. Google discovered that blue-ish links encouraged more clicking than greenish — so the search giant stuck with blue-ish. It’s only natural for Google to want you to click more: The more links you hit, the more time you spend online, hopefully looking at Google’s ads. As for why people prefer bluer links? I suspect they’ve come to expect them to be blue. So, when confronted with a greener link, they are less likely to click it. That’s just my unscientific theory, though — what do you think? And which color would make you click more?

You’re subscribed! If you like, you can update your settings

  1. Varun Mahajan Thursday, July 9, 2009

    I read somewhere that we like blue colo(u)r because we are naturally at-ease with colo(u)r of sky

  2. Design schools teach graphic designers how to direct the user’s eye with hierarchies of color (from high contrast to low), fonts (from large and bold to small and light), and element position & flow. When most people see a page of links, there eyes will go first to the highest contrast (“brightest”) links, all else being equal. My guess is the blue links simply attract the eye better than the green links, because they stand out more.

  3. So if Google has an answer from the testing, why are the link colors still different between Gmail and Google’s main results? Which color “won” and resulted in more clicks?

    1. Jordan Golson Jay Thursday, July 9, 2009

      I think the different shade of blue on Gmail was a design decision from the start — the test they were doing was between greenish links and blueish links — the color that Gmail is using now may be the one that won.

      I would be surprised, however, if Google were to change the color of their search results pages. Those are the company’s bread and butter — if it ain’t broke…?

  4. This behavior caused Google to lose one of the best interface designers in the world (Doug Bowman), and is indicative of a deeper problem in the culture of the big G — an over-reliance on data, and an under-reliance on expertise outside of engineering.

    If you want the best color of blue for your links, that color of blue should be a pure design decision, made by someone with strong expertise in design, psychology, and human-computer interaction. It should not be something that’s opened to a (unintentional) vote among your users.

    1. Not if the designer’s decision costs you million of dollars. Think about this from a business perpective – you can have one of the best designers on Earth working for you… or make an extra $10m (figure pulled out of the air). As a business, do you take the designer or the ten million? Right. You take the $10m. If it’s a much larger figure, the decision just gets easier. Design in a business has to serve the business – if it’s merely serving the designer, something’s off.

      Google doesn’t make their money on awesome design, they make it on design that converts people – which for Google means clicking links.

      1. Are you kidding? Gmail finally gave up old, stale simplicity and adopted new themes for no reason? Had it left on engineers to decide; they would have been toast. UI experience is an art too. There were MP3 players and slick smart phones before Apple entered the market. Had it taken google approach; it would have never been successful. Google search is relevant for now because of its brand advantage over other search engines. If you ask someone who is not in tech business to name a search engine other than google; it would be difficult for him/her. Look at how google news and finance pages look and compare them to yahoo news and finance. Saying that elegant design doesn’t matter is the stupidest thing.

  5. A couple comments about the article and the comments.

    First, you’re a guinea pig all the time. Large companies test market their products in super markets all the time. They release a product in one market to see how consumers like it. They change the prices or offer coupons in specific markets to see how it impacts sales. I’m sure they do design changes as well. One of the advantages of being a mass consumer product company is that you can run trials like this.

    I’m not really sure the designer who quit is really a Google issue at all. Yes, design is nice, but it shouldn’t win out over more effective systems. Google is a business not an art foundation. If they wanted nicer design they could have easily made their home page much heavier. It would be pretty but from their research so far, it would make users less efficient. That may make the designer unhappy, because it isn’t a fulfilling job making art, but that just means he should be working somewhere else.

  6. @Tox

    It could be simply a culture issue in how the experiment was raised/handled. If the experiment was used as an opportunity for the ID designers at Google to author a paper or publish an interesting result in scientific literature — eg. provide data to back up the ID textbooks and classes. the key is to provide appropriate motivation for the different folks/disciplines involved on the dev team, imo

    —– Tox wrote ——-
    This behavior caused Google to lose one of the best interface designers in the world (Doug Bowman), and is indicative of a deeper problem in the culture of the big G — an over-reliance on data, and an under-reliance on expertise outside of engineering.

    If you want the best color of blue for your links, that color of blue should be a pure design decision, made by someone with strong expertise in design, psychology, and human-computer interaction. It should not be something that’s opened to a (unintentional) vote among your users.

  7. art vs commerce, Microsoft won over Apple, goes to show technical triumph over design. Google knows that, what is the point of having fancy design but lousy search results?

    1. They did? Look at the stock prices of the 2 companies over the last decade. Look at the balance sheets. MS is *bigger* but both companies have won and in stockholder eyes over the last 10 years, Apple is the clear winner.

      1. Stock prices? Apple was in a bad place when Jobs took back the reins. Since then, it has done quite well, especially compared to Microsoft, but that’s like comparing oranges to, well, apples. Even 10 years ago, Microsoft was an established company, signaled by its addition to the Dow. Apple was a struggling, mismanaged company. It made quite a comeback–mostly thanks to the iPod–but it had far more up-side than down-side potential. Microsoft, meanwhile, has been uneventful–exactly what’s expected from a Dow component.

        There are a few telling things going on right now. MSFT outperformed AAPL in the past year; this might be a sign that Apple’s growth is peaking. APPL’s market cap is greater than newly-added Dow component CSCO, but its net income (profit) is less than both CSCO and MSFT. AAPL is still priced for growth, but its performance over the last year was unimpressive.

    2. are you seriously implying that Apple = Art and Microsoft = Commerce? or did you mean it the other way around?

  8. This is all about the sub-conscious. HTML links were originally blue, so we’ve become conditioned to clicking on such blue links.

  9. electricdave Thursday, July 9, 2009

    Blue is stupid – there I’ve said it. Maybe I have the design sense of a rat from the Seattle landfill, but green is clearly the most important color. Why? Because I said so. Don’t need statistics from millions of users to prove me otherwise. *snark*

    (I read that article about Bowman and thought he was bitter. He should have found a way to incorporate a sort of crowd-sourced design methodology so it matched the statistical methods in use. Then he would be a genius. Good designers are like gold, but you have to appreciate that speed trumps everything when it comes to clicks.)

  10. Trouble with this is all monitors display colors slightly differently – unless one uses calibrated only – so one doesn’t really know what color people actually saw. But somewhat depends on the size of the test, do the results translate to the population as a whole or just those tested. And then there’s color blindness in the visitors.
    To a big degree it’s Google engineers mentally masturbating.
    ———-
    In all my designer years, I’ve never met a male client that didn’t prefer blue. Design a logo, make it blue, sold. Woman have been all over the board.

  11. links for 2009-07-09 « Jet Grrl Thursday, July 9, 2009

    [...] When It Comes to Links, Color Matters (tags: interwebs WebDesign) [...]

  12. Well, people click on blue links because we grew up with it. Had it been green from the beginning; the tendency to click on green would have been higher. The monitors in late 90’s weren’t great at displaying various colors thus, it made sense back then to have blue links which could be easily distinguished.

  13. Libran Lover Thursday, July 9, 2009

    And the point of this post being…?

    I read that the paid version Gigaom Pro was supposed to have more in-depth articles. Didn’t realize that the free version would simultaneously move to shallow waters.

  14. Google’s Color Test on Click Through Rates of Links | YieldBuild Blog Friday, July 10, 2009

    [...] the click through rates of links.  Google did a recent expirment with GMail that was covered by GigaOm on the click through rates of links by altering the link color. But there’s also the fact [...]

  15. well the “good old” blue link color has found it’s place deep in any computer users mind, it is obviously the “natural” choice to click on blue links…

  16. links for 2009-08-30 « links and tweets Sunday, August 30, 2009

    [...] When It Comes to Links, Color Matters When It Comes to Links, Color Matters http://ff.im/-51LXh [from http://twitter.com/kenmat/statuses/2560637430%5D (tags: tweecious Google Microsoft Apple Websearchengine search SEARCHENGINE Opensource Gmail) [...]

  17. Just look at the link color on craigslist…hmmmm…clickity clackity…

  18. Das Google-Mißverständnis – Die wunderbare Welt von Isotopp via @koehntopp « urban-listening Sunday, November 8, 2009

    [...] harte Daten haben kann – Marissa Mayer zum Beispiel ist bekannt dafür, daß sie einmal die Klickraten von 40 verschiedenen Blautönen hat Messen [...]

  19. All links of the early web were blue – perhaps it could be one reason ?!

    I like links in orange or red colour most and I think I click more often orange links than blue links.

  20. That nuance of any color really matters. As a color consultant, this is at the heart of my work – whether it’s the right shade of dark blue for the next generation of Xerox printers, or a clean green for a bank. See my new blog post at http://tinyurl.com/ye6ndmo
    (http://colormatters.blogspot.com)

  21. Google have started using ‘pukey blue’ on some of the links in the redesigned search results page.

    Luckily with Firefox it can be changed by tweaking userContent.css.

    I much prefer keeping unvisited links a consistent colour, #2200CC. Is that good design or not? – don’t know but that’s how I like it!

  22. @gisardo Google has often done this; test small changes on subsets of users http://t.co/U819Ip5r I don’t approve of chrome only stuff though

Comments have been disabled for this post