How Link Color Affects Conversion – Split Test Results

PLEASE SHARE

Have you ever wondered if link color make a difference to conversion and engagement with your site?

As your resident sad act, I have!

Earlier in the month I took a look at if it’s better to have your sidebar on the left or right and whether testing logo designs makes a difference.

But what about link color?

Color Influences People

I’m not going to bore you with the ins and outs of color theory although it is an interesting topic.

But take a quick look at how these brands apply color-

color changes behaviour

Color can have a huge impact on behavior and as such it directly influences conversion.

This was seen when I changed the color of a button to increase profits.

As a general rule of thumb any calls to action should stand out from the rest of the page.

If you have a white/blue design – your calls to action should be red so they attract attention.

Things that stand out get noticed. People click on things that stand out.

Pink Links vs Red Links vs Blue Link Color

When I setup the blog I changed the color of the links to pink and headers to blue.

This was so the links stand out from everything else on the page as I explained above.

So the blog was already optimised to some degree in that respect, but can I take it further?

Setting Up The Test

I changed the color of all links within blog post content using the instructions in this post.

Link colors in the sidebar etc. were not changed.

There were 3 possibilities-

  1. Pink – This is the original link color
  2. Blue – This is the ‘classic’ blue link color we are trained to click
  3. Red – This is the ‘default’ #ff0000 red

The test ran from May 31st – June 12th and saw a total of 7,845 unique visitors between the 3 variations.

The Results Are In

I decided to look at how the change in link color affected key metrics such as-

  1. Bounce rate
  2. Visit duration
  3. Pages per visit
  4. Conversion (email subscription/affiliate click/resource downloads)

I also decided to segment the data in 2 ways-

  1. All Visitors – A look at all traffic both new & returning
  2. New Visitors – Just new visitors – it is likely that returning visitors are already ‘trained’ to use the site. New visitors provide a ‘fresh’ look

You can click on any of the images below to get the full versions.

So first of all let’s take a look at how it affected all visitors.

All Visitors

Bounce rate, visit duration & pages per visit are up first-

initial engagement test results

And then conversion…

initial conversion rate test results

Let’s break it down in table format to make it a bit easier to read, digest & compare.

Link Color Pages/Visit Visit Duration Bounce Rate Goal Conversion
Pink 2.65 00:05:53 47.82% 9.32%
Blue 2.65 00:06:05 47.44% 10.22%
Red 2.51 00:05:41 49.61% 9.84%

The blue links delivered the longest visit duration, lowest bounce rate and the highest goal conversion.

Seems like Microsoft did a good job of subliminally training us over the years with the classic blue link color.

Interestingly the red links performed pretty poorly but stood out the most against the blue/pink theme of the site. More on why that is later.

Perhaps visitors that haven’t seen the site before will behave differently.

New Visitors

Bounce rate, visit duration & pages per visit are up first-

link color general test results

And then conversion…

link color conversion rate test results

Let’s break it down in table format to make it a bit easier to read, digest & compare.

Link Color Pages/Visit Visit Duration Bounce Rate Goal Conversion
Pink 2.23 00:03:45 53.13% 8.09%
Blue 2.14 00:03:41 53.55% 8.85%
Red 2.23 00:04:00 52.57% 9.56%

When we look at just new visitors things are very different.

This time the red links are the best performing in terms of visit duration, bounce rate and goal conversion.

The pink and blue links perform at similar rates with the blue links edging forward on goal conversion.

But why did this happen?

Wrapping It Up

When both new and returning visitors were shown red links it had a negative impact when compared with the pink links.

But when just new visitors were shown red links it had a positive impact.

Why though?

Returning Visitors

I believe this is because my returning visitors are-

  1. Already engaged with the site
  2. Already familiar with the original pink link color/brand

The red links were also pretty ugly and I felt like they decreased readability, so this was a ‘step back’ for my loyal readers.

My regular readers are already ‘trained’ to click pink links-

Facebook link color feedback

However they couldn’t overcome the subliminal blue link color that Microsoft has trained us to click on over the years.

The blue link text color actually blended into the branding of the site better than the red links while still standing out as clear calls to action

New Visitors

New visitors behaved differently because they didn’t have the pre-conceived familiarity with the site and were seeing it fresh for the first time.

They didn’t have a benchmark of design or readability to fall back on like returning visitors.

At the start of this post I said that “points of conversion should stand out like a sore thumb” and the red links certainly did that.

They were upfront and in your face to the point of distraction but this wasn’t a step back for new visitors as it was their first step.

So how on earth am I meant to pick a default link color winner?

ALERT: Instantly Increase Your Conversion Rate With My Personal Case Studys!

Can You Retrain Visitors?

Based on the data the logical thing to do is show new visitors red links and returning visitors the blue link color.

But that doesn’t seem very practical and will lead to user confusion long term.

So that has me wondering…. Is it possible to reprogram my loyal readers to show the red links some love?

If I just forced the change on you guys you would have no choice and over time you would be ‘retrained’.

That’s what Facebook do when they want to change things up right? Although that always goes down like a knackered lift for a few days.

What color do you think the links should be? Answers in the comments please – your feedback decides this one!

69 Responses

  1. 6.21.2013

    For me Pink is nice, not ugly and not aggressive,
    blue for me can mean this is again some spam link or something wastefull

  2. alexleigh
    6.21.2013

    Woodward! Stop distracting me from work with your compelling content

    • June 21st, 2013 at 4:09 pm

      Shut up and read it :P

      • alexleigh
        June 21st, 2013 at 4:11 pm

        Rude.

        :-)

        • June 21st, 2013 at 4:11 pm

          Haven’t you got work to do? ^^

          • alexleigh
            June 21st, 2013 at 4:18 pm

            I did, but I keep having to come back here and drop comments.

          • June 24th, 2013 at 9:16 am

            Monday mornings are always busy though right :P

          • alexleigh
            June 24th, 2013 at 3:02 pm

            They should be…

          • June 25th, 2013 at 8:45 am

            I dont know you know but I do know that you dont know you know.

            Try saying that as fast as you can

          • alexleigh
            June 25th, 2013 at 9:08 am

            That sentence feels like it needs to be read in a Norfolk accent

          • June 29th, 2013 at 9:00 am

            Hahaha – it does actually mean something though ^^

          • June 22nd, 2013 at 10:05 am

            Matthew, you are priceless (and brilliant!)

          • June 24th, 2013 at 9:06 am

            Haha thanks

  3. 6.21.2013

    Interesting stuff. I guess there are endless combinations you could try: underline, hover styles, etc

    Is there a way to determine if a user is new or not and then show them links/content in a different way?

    Great case study though, keep them coming!

    • June 21st, 2013 at 4:08 pm

      Yeah you could place a cookie and load a different stlyesheet based on that.

  4. 6.21.2013

    Wow. This is pretty impressive. Very nice read.

  5. 6.21.2013

    Hi,

    great post really. It’s interesting that red always outperforms in terms of conversions.

    To my mind: Use Red! Why? Because over time your “old” visitors will get used to it and the new visitors will show higher conversions anyway. Think long-term, in the beginning you would lose some $, but I would say that after a certain period of time everyone will love red.
    The question is: When :)

    • June 24th, 2013 at 9:17 am

      Hi,

      I wouldn’t say it ‘always’ outperforms but as a rule of thumb its a performance colour – still needs to be tested though.

      So you think I can retrain my visitors :)

      • July 1st, 2013 at 3:46 pm

        Yes, of course. Humans are experts in habituation.

        Think of all other well-known web-projects. How often they have changed not only colors but the complete design and layout.

        This happened to fiverr lately, I still cannot accept the new design :) As you see it takes time…

        • July 3rd, 2013 at 10:21 am

          Great feedback and a nice observation – its just a matter of time :)

  6. doof155
    6.21.2013

    I think you should use red if it works best for new visitors.

    Red may not work better with the ‘regulars’ but as you say it is something that people will get used to and as long as you are still providing the quality content then it certainly won’t steer people away from clicking I wouldn’t imagine!

    Another great ready by the way Matt!

    • June 24th, 2013 at 9:16 am

      Hi,

      Thanks very much – so basically f*** em and do what I want huh :P

  7. Hector
    6.21.2013

    How do you set up these test…it would be great if you can show how to set it up in analytics. I believe it would make a great tutorial too.

  8. 6.21.2013

    I’m ALL for blue though… all shades of..
    Always outperformed in my tests – as trust is a BIG item in the equation of persuasion…

    I know there are other arenas where pink would be Queen-of-the-Hill though… heh… go figure!

    Cheers,
    Steve ✉ Master eMailSmith ✉ Lorenzo
    Chief Editor, eMail Tips Daily Newsletter

    • June 24th, 2013 at 9:13 am

      Hi,

      Have you found that to be the case across different niches/audience types?

  9. 6.21.2013

    I really like the pink links. With the blue headers and white background I think it stands out without being too much of an eyesore. Red would probably decrease readability for me, and blue is just downright unoriginal :)
    Being that you even supplement your link text color with an underline, I don’t think you have anything to worry about.

    And I fear change!!! No change!!! BAD AAHHH!!

    • June 24th, 2013 at 9:10 am

      Good points and what you say about underlining is probably true – I think this is something we are all ‘trained’ to look for with a link

  10. Paschalis I.
    6.21.2013

    Well the returning visitors already know the value of their site so I don’t think a color change will harm them so much. We will get used to it.

    So do what the conversions say it’s the best for you. In the end everybody will get use it, and you’ll have less new visitors to retrain if you decide to do it later.

    the sooner the better I suppose ^^

    that’s my opinion,
    good luck

  11. ogenox
    6.21.2013

    Pink, pink, but only for women related sites. :-p

    • June 24th, 2013 at 9:08 am

      Hahahaha I think the pink my site is rather fetching if I’m honest!

  12. Ian Wilcox
    6.21.2013

    Oh Matt you have far too much time on your hands :)

    I like the default blue links, i think that most people (Well in the UK anyway) know that when text is in blue its a link.

    Maybe you should throw putting the link in bold, italic and underlined into the test :-p

    • June 21st, 2013 at 3:58 pm

      Hahaha I dont have time on my hands – thats why its critical to build a finely tuned conversion machine ^^

  13. Charles Floate
    6.21.2013

    I found myself wasting 10 minutes just looking at your color emotion guide (which should be colour, your on a .co.UK remember) anyways the moral of this story is:
    PRETTY COLOURS!

    • June 24th, 2013 at 9:14 am

      My grammar/spelling isn’t great at the best of times – you’ll also see me make mistakes with been and being on a regular basis :)

  14. 6.22.2013

    I see from the diagram at the top that Hooters are in orange, so that’s good enough for me!

  15. 6.22.2013

    Again – great report! I was waiting for this post – I had noticed your links changing colors…knew something was coming!

    I also like that “Color Emotion Guide” – I had never seen anything like that before. Good stuff to know!

    • June 24th, 2013 at 9:07 am

      I didn’t go into the depths of colour theory in the post but check it out, its a very interesting subject!

  16. 6.22.2013

    Thanks so much for sharing all this info. So much to learn about about my “new online biz”

  17. 6.23.2013

    Hi Matthew,

    I’m rather new to split testing so your articles have helped me understand more how it’s done. I came across something called “statistically significant” which, to my understanding, help to decide whether we can make conclusive decision based on the result.

    I’m using the calculator provided by Visual Website Optimizer at http://visualwebsiteoptimizer.com/ab-split-significance-calculator/

    I don’t see the actual number of visitors from the screenshots so I calculate based on ‘Goal Completion’ and ‘Goal Conversion Rate’. For example for new visitors, the original colour has 168 conversions at 8.09%, which means you have around 2076 visitors who saw the pink links.

    If I plug the numbers into the calculator, the result comes out as NOT statistically significant.

    Is this the right way of interpreting the result?

    • June 24th, 2013 at 8:59 am

      Hi,

      I haven’t used that tool but there were enough visitors exposed to each test to make a decision. Generally a test of 1,000 visitors per variation is enough to give an indication.

  18. 6.23.2013

    Another great case study! It’s interesting to find how the pink and blue links have a diff. impact on diff. visitors (such as new and returning). I’m actually about to run a split test between a plan white background on one of my blogs, and a background image. I’ll be tracking the same metrics as you did.

    • June 24th, 2013 at 8:57 am

      Sounds good let me know how it goes! Interesting test!

  19. [email protected]
    1.3.2014

    Mr. Woodward. You have this down to a science. Great post. Keep’em coming.

  20. 5.11.2014

    Very helpful information and makes me feel good as I have always used blue for link color, now I know why. Will be looking
    for more from you it seems you have some good knowledge on these matters. Thanks

  21. 10.29.2014

    This is an awesome case study, but I really think you’re running it in the wrong niche. I would be more interested to see how a “regular” Internet user reacts to the different color links (someone shopping for a nice Christmas scarf or something). I would almost bet money that the classic blue link would perform better with a different audience. Your audience of savvy IMers is probably not a good representation of the typical user.

    Oh, and gingers have no soul.

    • October 29th, 2014 at 3:32 pm

      Hi,

      Yes I agree and would love to run it in another soon :)

  22. 1.21.2015

    Blue links because – blue is seen as dependable?

    Is it possible to have orange links?

    • January 22nd, 2015 at 3:13 pm

      Blue because thats what colour links have been since the dawn of internet explorer.

      You can have any colour you want

  23. 2.24.2015

    Matt, this is the first article of yours I have ever read — found it after Goolging “hyperlink colour”.

    Great study, great data, great write-up, great rationale. You have just converted a new visitor to a repeat one.

    • February 24th, 2015 at 8:24 pm

      That is great to hear Michael thank you for your readership!

  24. 12.30.2015

    red links all the way!

  25. 4.4.2016

    Guess you stayed with pink then; love to know why?

  26. 4.27.2016

    I like blue links but I’m old school guy. I’m trying different colour of links, especially in WordPress articles and also trying to find out which works best.
    The other matter is interesting here though – backgrounds. Have you tried also with different text backgrounds? That would be interesting.

    • April 28th, 2016 at 9:04 am

      No I haven’t played with backgrounds myself because I think it looks spammy on text but that is just an opinion :)

  27. 5.18.2016

    Great Article about the link colour should be!!! before the read, I never considered link colour but I used to give the light blue colour to my Post links just because it highlights a phrase or combination phrase perfectly, Now I am getting know-how link Colour affects on clicks rate.
    Thanks Matthew for your Nice Concern.

    • May 19th, 2016 at 11:42 am

      Basically, as long as the link colour stands out and doesn’t blend in with other colours on your site – you are good!

  28. 10.9.2016

    Just came across this article which is pretty inline to what I was testing on one of my blogs. And I found that blue colored links are clicked73% more than the ones without any color.

    Although your tests were quite extensive than mine. I just tested one color.

    Anyways I always pick gold nuggets from your articles. And the thing I love the most about your posts is that they are updated on regular basis, no matter how old the post is. Which is quite rare in our industry.

    • October 9th, 2016 at 6:12 pm

      Thanks Dru – glad you found it useful and conducted your own tests as well :)

What are your thoughts?

* Name, Email, Comment are Required