Tag Archives: Web analytics

:: Standard Metrics Revisited: #3: Bounce Rate

12 mar



6 August 2007

When Ann asked me to start writing for Marketing Profs Daily Fix I was all excited and quickly wrote a article on a web metric that has enamored me recently, bounce rate. Unfortunately, as readers of this blog are well aware, I wrote too much. The nice thing is now weeks later I can post the whole thing here!!

This post is a expanded version of the Marketing Profs article (link: Bounce Rate: Sexiest Web Metric Ever?). It is updated, it has two additional best practices and more context at the end.

[For those of you who attended the all day Google Conversion University web analytics education seminar there is lots more detail here – there was so much more I wanted to cover last Wednesday!]

It is quite likely that your company is spending tons of time, energy, and dollars on web marketing efforts yet conversion rates (or ROI) are stuck in the two to three percent range.

You are trying really hard to figure out how to improve the performance but you are stymied by the fact that there is ton of data and you have no idea where to start.

Ms. Bounce Rate to the rescue!

Bounce rate is a beautiful way to measure the quality of traffic coming to your website. It is almost instantly accessible in any web analytics tool. It is easy to understand, hard to mis-understand and can be applied to any of your efforts.

So what is this mysterious metric? In a nutshell bounce rate measures the percentage of people who come to your website and leave “instantly”.

Thought about from a customer perspective rather than I came, I saw, I conquered, the action is I came, I saw, Yuck, I am out of here.

Your marketing efforts should yield more customers who are able to conquer (accomplish the task they are there to accomplish) and fewer who say yuck and leave.

Bounce rate measure quality of traffic you are acquiring, and if it is the right traffic then it helps you hone in on where/how your website is failing your website visitors.


It is usually measured in two ways:

  • The percentage of website visitors who see just one page on your site.
  • The percentage of website visitors who stay on the site for a small amount of time (usually five seconds or less).

I am a bit more aggressive in my analysis and so I personally prefer the latter definition (using time). But either definition is fine, each has its own slight nuance. Please check what your tool’s definition is and make sure you understand it and communicate it to your data consumers.

So how can you use it?

#1: Measure the bounce rate of your website.

This will help you understand what percent of your website traffic is actually engaging with your site (i.e. you don’t have a chance of success with all your site traffic!).

indextools time trend bounce

This IndexTools report shows that approximately 70% of the site traffic above is bouncing. This 70% won’t even give you five seconds, or see more than one page. You can’t even start to impress them with your goodness. No matter how great your goodness is.

There are two exceptions: 1) You have a one page website 2) Your offline value proposition is so compelling that people would see just one single webpage and get all the information they need and leave.

If you don’t fall into those two categories you need to pay very careful attention to this metric.

Action: You’ll understand better why your conversion rate is so low, if you have made changes over the last x amount of time then watching a trend of bounce rate is a sure way to know if the changes you are making are for the better.

#2: Measure the bounce rate for your traffic sources.

Your goal is to figure out if some sources of traffic are sending you particularly terrible traffic compared to others. In your web analytics tool simply go to the Referring URL’s / Sites report and take a look, you’ll find something like this……

google analytics referring sites bounce rate

For this site Google Analytics illustrates that both myspace.com and simplyhired.com is not sending great traffic, while their direct marketing campaigns (#2 and #3 above) seem to be doing much better.

Action: Do you need to revisit relationships with sites that are not sending you high quality traffic? What is the call to action that is causing people to come to your site and bounce? Are your email, affiliate, other marketing campaigns yielding low bounce rates? You get the idea.

#3: Measure the bounce rate of your search keyword.

Search engines are king and hence probably sending you a large chunk of your traffic. Find out the quality of traffic coming from the search engines. In addition to looking at just the aggregate levels (the row marked Total in below report) also look at each key phrase……

clicktracks search keywords bounce rate

In this ClickTracks report notice that each key word / key phrase performs differently, and it also differs by search engine. Sweet actionable information. Instantly useful!

Action: When you create SEO and SEM campaigns ensure that your team has this data. In this case either you have nothing to do with competitive intelligence (hence 90% bounce rate) or that traffic is landing on the wrong pages. Also it is easy to take a dump of top keywords and bid on them with PPC campaigns, the table above will ensure you don’t bid on the wrong ones.

#4: Measure bounce rate of your AdWords, AdCenter, YSM (PPC) campaigns.

In my humble experience this is one piece of analysis most agencies and companies overlook when doing paid search campaign analysis.

Sure we measure conversion and roi and revenue, but are you measuring bounce rate for your PPC campaigns? Remember you can only convert if people are staying for more than five seconds on your website (or see more than one page)!

google analytics adwords bounce rate

This screenshot from Google Analytics shows the bounce rate of traffic on each keyword compared to site average, very cool view. Sadly most traffic for this time period is performing worse than site average (so literally you could be sending money down the, well you know what).

Action: First, stop bidding on those keywords, then do a deeper analysis of how good your landing pages are, and your other campaign attributes (maybe your campaign for refrigerators is being targeted to people only in the great state of Alaska!).

#5: Measure bounce rate of your top trafficked pages.

It is entirely possible that your efforts are stellar (as they usually are!) but it is your website that is letting you down. There is what to do to make your case…..

google analytics content bounce rate

What pages are bouncing traffic like a perfectly formed elastic material and which are great at welcoming traffic with open arms into your website? Pull up the above report in your web analytics tool and find out.

In addition to top trafficked pages on your site please also look at top entry pages report and measure their bounce rate (you’ll quickly have a list of pages that need your immediate attention).

Action: Check to see if the right calls to action are on the page? Is the content optimally organized? If the above pages are your campaign (direct marketing or paid search campaigns) landing pages then are they delivering on the promise of the email piece you had sent out or the search keyword? Answer these questions and consider multivariate testing to improve page performance.

An Exception :

There is one obvious case where bounce rate might not cough up as many insights. I am thinking of blogs.

They are a unique beast amongst online experiences: people come mostly only to read your latest post, they’ll read it and then they’ll leave. Your bounce rates will be high because of how that metric is computed, and in this scenario that is ok.

You don’t want the bounce rate to be 98%, new visitors to your blog will still come and look around and read different posts etc. But I would not worry if my bounce rate for this blog is 50%. Maybe at 75%. 🙂

In summary:

  • Bounce rate is a metric you’ll easily find in all web analytics tools.
  • It is easy to understand, hard to misunderstand (something you can’t say for all, or even most, web analytics metrics).
  • It is awesome at identifying low hanging fruit (you stress about so much, start here and you’ll easily find so much to action), try the above five tips first.
  • Please don’t confuse bounce rate with exit rate, they are radically different metrics. Also: Everyone who comes to your site has to exit, almost no one who comes to your site needs to bounce!

Would you agree this is a awesome metric? It won’t have all the answers for you, but it will help you focus very quickly on what’s important, show where you are wasting money and what content on your site needs revisiting.

As a benchmark from my own personal experience over the years it is hard to get a bounce rate under 20%. Anything over 35% is a cause for concern and anything above 50% is worrying.

What do you all think? Please share your tips, tricks, war stories, critique, brickbats via comments.

[Like this post? For more posts like this please click here, if it might be of interest please check out my book: Web Analytics: An Hour A Day.]


:: Aggregation of Marginal Gains

11 mar



10 March 2009


We are constantly on a quest to conquer the next big thing. Mountain. Ocean. Planet. “Conversion Buster.” The next million dollar opportunity.

Not that there’s anything fundamentally wrong with that quest.

The challenge is that frequently in that quest we ignore the immediately achievable. And that tradeoff is a crime.

The title of this post comes from a blog post that my friend David Hughes had written in September 2008: What can Digital Marketers learn from Olympic Cyclists? He started that post with this wonderful quote:

Back in the 1980’s Jan Carlzon was trying to breathe new life into an ailing Scandinavian Air Services. He was famous for saying “You cannot improve one thing by 1000% but you can improve 1000 little things by 1%”.

Aggregation of marginal gains!

Fantastic concept, loved it instantly, captured my heart.

This post is all about the low hanging fruits. Small and medium sized ideas for finding opportunities that collectively should add up to something remarkable for your website.

This being the recession and what not, please share your own stories of simple every day things you do to find actionable insights for your company. Together we can! : )

So before your go boiling the ocean here are a few things you can and should do today to ensure your company is benefiting from immediately fixable things:

1) Figure out where you are making money, where you ain’t.
2) Cover all the bases in your email campaigns.
3) Funnels baby!
4) Stop The “Puking”: Fix Your Top Landing / Entry Pages.
5) Identify Paid Search Keyword Opportunities.
6) When In Doubt, Ask Your Customers.
7) Stop Doing “Dumb” Things. (Examples included!)

Batten down the hatches, let’s deep dive…

1) Figure out where you are making money, where you ain’t.

Why is it great?

You come to work to:
a) Increase Revenue
b) Reduce Costs
c) Improve Customer Satisfaction (create Brand Evangelists).

This analysis focuses on the first two of those things.

It helps you identify which sources are working well (and hence need more love, resulting in increased revenue.

It will help you identify sources that might be sucking wind, and help you reduce cost.

Possible outcomes?

Focus in your advertising, sales, marketing spend. Identification of valuable sources that might have been below the radar so far (#2 or #10 for me above).

2) Cover all the bases in your email campaigns.

This one comes from David’s post.

All of us run email campaigns, most of us professionally (and we live in Russia and sell spicy things!). But have you covered all the simple basic things you should doing to maximize value?


He recommends that by fixing five simple things he was able to really crank up the conversion rate for a seminar, by 486%.

[Of course having penned the Four Not Useful KPI Measurement Techniques post I hasten to add that improvement was from 7 to 41. Now before you roll your eyes consider this, would you not kill to have a 20% improvement in your email campaigns?]

Why is it great?

It is achievable.

Each thing he recommends is quite simple. For example fix the Hard Bounce Rate by using email repair software, or Soft Bounce Rate by identifying spam blocks, or Open Rate by testing subject fields (or, my fav, optimizing for the Preview Pane), or creating illusions of personalization for Click Thru Rates, or continuing the “scent” to landing pages to improving conversion.

Small things that add up to a lot.

Possible outcomes?

Email continues to be one of the most lucrative channels for most businesses (even spammers!). Doing things above means you make a better connection with your customers (that itself is worth it) and improve Revenue.

3) Funnels baby!

I rarely talk about funnels (specifically cart/checkout or credit card application or lead submission). That’s because I think everyone’s doing ‘em.

But it turns out I am wrong.

A small fraction of the people use funnels and that’s a real shame because you can find actionable things to fix pretty fast.

You need a funnel report in your web analytics tool! Demand / command that you have a funnel created for each structured experience on your site. Ok beg if that is what it takes.


They look so pretty too don’t they?

Why is it great?

For structured experiences funnels can be awesomely actionable. They can identify problems quickly, is 73% abandonment on step one acceptable (seems a bit harsh I think) or how about 75% on step two.

And talk about low hanging fruit. At the end of the day you are dealing with one page. One. Uno. Ek.

You know why that page exists. You know where people come to it from (look at the box on the left). You know where people exit to (look on the right). Not that hard to come up with two or three simple fixes to put in place (or A/B tests to get going with the Google Website Optimizer!).

Possible outcomes?

Most structured experience on sites are money making / saving opportunities. Shopping Carts (or Baskets to my British friends). Lead requests. Applications. What not.

You will rarely get this close to having a direct impact on the bottom line.

Initially you’ll be ok without segmentation but later you’ll learn that segmentation is your BFF. It is hard to do this with Google Analytics (you can apply the filters and profiles magic) but it is pretty easy to do it in a tool like ClickTracks . Do it.

4) Stop The “Puking”: Fix Your Top Landing / Entry Pages.

Another eternal favorite of mine when I want to get some quick wins under my belt. It is one of the first things I look at when I analyze any website.

Most web analytics tools have a standard reports that show the Top Landing Pages by Bounce Rate. Go get it.

Yahoo! Web Analytics (IndexTools) does not have a standard report but click on Customize Report and in two seconds you have:


Note what I am doing with the stars: looking for landing pages which really high bounce rates. That’s what you’ll do on your site, find the pages that are “losers” who can’t even earn one click.

[Sidebar: if you look at bounce rates for blogs remember it is not the best metric, unless you segment out new visitors in which case it is still helpful. Read how to analyze bounce rates for more.]

Why is it great?

Like we did before, you are looking at high value individual pages. Lots of traffic, lots of bouncing.

You already know why the page exists (just read it!). And when you look at these pages it is easy to get to the thing you covet the most: Customer Intent!

Check two important things (both available on the same report): External sources referring traffic to this page (”why?”). External keywords referring traffic to this page (”what do they want?”).


But give you a quick understanding of what the mismatch might be between your intent on the page (sell sexy pants) and what the customer wants (v1agra!).

If you want to go one step deeper go to your internal site search report and look for keywords people in your internal search engine on that page.

So on your sexy pants page people are searching for sexy pants something’s wrong with the page.

Possible outcomes?

Increased possibility of conversions / leads / watching videos / whatever the purpose of your main pages are.

And remember these are your “head” pages, small improvements yield high ROI.

5) Identify Paid Search Keyword Opportunities.

Search, organic or paid or both, tends to be a corner stone of the acquisition strategy for many companies. Lots of gains can be gotten from wise investments in search.

One beautiful actionable idea for you. Do a diff between your paid and organic search keywords and look for anomalies.

Keywords that have a better conversion / time on site / bounce rate / your fav KPI under the Organic bucket when compared to their performance in the Paid search bucket.

Why is it great?

If you find keyword that perform better when they bring Organic Search traffic then investigate the landing pages and see what is better / different about them then the Paid Search landing pages for that traffic.

Compare the organic listing in Google (ok, ok, or in Live or Yahoo!) with the Paid Search ad. Is there any different (better) about the text and call to action in the Organic listing?

In both cases learn from Organic and fix in Paid. If the Paid performance is truly sucky then rethink PPC for those keywords.

Do the reverse for Paid Search. Look at the above report and see where are places where your PPC campaigns are doing better. Why? What can you learn and implement on the Organic results pages?

Often there are lots of learnings from the Paid campaigns because your landing pages are much simpler and more targeted than your Organic pages. Take the lessons and apply them to the other.

Possible outcomes?

Search is a complex and tough thing to do right. Covering the basics above means that your company will start to be a learning machine where you are not discarding any scraps of information you are collecting on your site.

At the minimum you’ll improve conversion rates. With just a little effort you’ll also reduce cost of Paid Search.

6) When In Doubt, Ask Your Customers.

How many times have you heard me talk about limitations of ClickStream data?

Far too often.

One more time: You can try to torture your data and guess. But why not ask your customers?

Use the free 4Q site level survey to get this key information:


“Dear Visitor, why are you here and pray tell us all the ways in which we have disappointed you?”

See? No guessing. Fix those things in Orange and you have a leg up on your competition.

More insights into Customer Intent! The nice thing is you can get some offline signals as well (”came from a tv ad”, “your radio campaigns were sweet” etc).

4Q’s for site level. If you want to improve individual pages then I recommend getting a page level survey to collect tactical can be quickly fixed information. Check out Kampyle and GetSatisfaction , both are wonderful.

Why is it great?

Why rely on your opinions and then fall flat on your face? Why try to argue with a HiPPO and let her decide what’s best for your customers?

By getting your customers involved means you magnificently increase the chances that you’ll get it right. More often.

Possible outcomes?

You will certainly increase Outcomes (conversions / leads / movie tickets / electric shocks to sub par customers, sorry I meant employees!).

But what is cooler is that you’ll focus on the 98% that will never convert and you’ll improve their task completion and satisfaction. So more offline conversions. More micro conversions! [See this post: Measure Macro AND Micro Conversions .]

7) Stop Doing “Dumb” Things.

Pardon my french, and please know I use that word with love.

It is hard to believe that in 2009 we still have websites that ask us every time we visit to pick a country (www.burgerking.com), even if I visit an hour later! How lame is that (!!). Or that have pages that look like this:


That is a page from a major content site where I have covered the navigation elements in blue and the ads in black. 8 ads.

In white is the actual content delivered to me. At 1440 screen resolution!

Astonishing right?

I call it the “selfish lover” strategy. Just think of yourself and what you can get out of it rather than the other person (customer!).

Have you ever loved a selfish lover back?

Why do you think it would work on your website?

If you have Web Analytics: An Hour A Day then on Page 59 you’ll find a Heuristic Evaluation checklist. Basic things you should do to make sure your have a good website experience.

Why is it great?

Do you really need me to tell you why? I am not going to insult you.

Except perhaps to say that before you pay thousands of dollars for to a external consultant / agency to “improve” your site, this simple checklist (or mine in the book) will yield huge benefits.

And they are all really simple easy to fix things.

Possible outcomes?

Happy customers. Engaging experience.

And a nice bonus, no one will call you out in their presentation / blog for being a poster child of how not to do website usability.

That’s it! End of story! Actionability awaits you!

Your turn now.

Please share your own ideas for “aggregation of marginal gains”. Is there a report / piece of analysis that always works for you? When you crack open Omniture or WebTrends or Google Analytics for a new site what do you look at first? Have a to die for KPI or Metric? What is one completely flawed assumption I have made in this post?

I would love to hear from you.


Avinash Kaushik