TopRank Marketing Editor

Website Analytics vs. The Myth of 100% Accuracy

Deep Dive Into Website Analytics At SES San FranciscoLet’s get this out of the way.  100% accuracy does not exist in website analytics.  Repeat. 100% accuracy does not exist in analytics.

What does exist in analytics is data – lots of it – and with this comes fear.

Fear of looking at the wrong data.  Fear of where to start.  Fear of analytics failure.

This fear will only dissipate with knowledge.  Knowing that analytics will never be perfect is a critical first step and a cornerstone shared during the session ‘Deep Dive Into Analytics’ at SES San Francisco.

Bryan Eisenberg, SES Advisory Board and NY Times bestselling author, moderated this session which included on its panel:

  • Tami Dalley, Director, User Experience Optimization, ROI Labs
  • Marty Weintraub, President, aimClear
  • Matthew Bailey, SES Advisory Board & President, Site Logic Marketing

The beauty of an analytics session at SES is that those on the panel may sometimes be analysts or they may sometimes be statisticians, but they are always marketers.  And a good marketing analyst knows that it’s the facts that tell and the stories that sell.

As such, Bailey shaped his presentation around four key points that make up the bulk head of the website analytics story:

  1. Intent
    Every visitor that lands on your website is there for a reason.  Bailey offers the example of a jewelry store pulling traffic for ‘watch’ based keywords.  Within this segment, some will be looking to ‘buy a rolex’ while some will be looking to ‘fix a watch + san francisco.’  Based on the keywords this segment used to find your site, did content match intent?
  2. Expectancy
    Similar to intent, but expanded to include referring sites ranging from forums to blogs.  Did the source that referred the segment, say a forum about watch repair stores in San Francisco, shape user segment expectations accurately in regards to what they found on your site?
  3. Reaction
    What ultimately happened when the segment of users arrived on your site?  Did they bounce immediately? Convert?  Did they do what you intended them to do?
  4. Behavior
    Whether the user bounced, exited or converted, what behavioral clues did they drop along the way? Did they bounce after a mere ten seconds? Exit after visiting multiple pages?  Convert directly from where they landed or within just one step?

Remove these items from a numbered list, and add context, and this becomes an actionable story:

Users arriving at my site using San Francisco watch repair related keywords, or coming from local websites geared towards watch repair services, are bouncing at a rate of 85% and converting at a rate of 4%.  However, those that land on my Rolex repair services page from this same segment bounce at a rate of 60% and convert at a rate of 10%.  If I optimize my Rolex repair page for ‘rolex repair + san francisco’, I may capture more conversions.

Of course, the answer is never quite so simple.  Dalley and Weintraub illustrate this as they share detailed real world case studies and conversion reports, respectively.

Dalley’s case studies detail how the slightest change, from a page’s design to a customer’s location, can drastically shape results.  The key to making an actionable change is to start with an educated hypothesis before backing it up with the right data for proof points.

And this means overcoming a fear of data and digging in, as repeated by Bailey.

Since Weintraub is sorely misrepresented in this post, primarily because nearly any written word would be a severe injustice to the energy level he brings to his presentation, below find potential conversion reports or data ‘dig-in points.’  These are reports that you can pull from your analytics to make a decision from today.  The list below was both leveraged from and inspired by a list of deeply ‘unsexy’ (Weintraub’s words, not mine) possible conversion reports shared during this session:

  1. Conversion by time of day
  2. Conversions by rural Pennsylvania city
  3. Conversions by mobile device
  4. Conversions by web browser
  5. Conversions by keywords containing the keyword ‘green dress’

Please add your own unsexy, or even sexy, conversion reports or analytics stories via a comment below.

PoorSo SoOKGoodAwesome (1 votes, average: 5.00 out of 5)


  1. Good article Mike. I have a question for you though; supposing you produce the conversions by time of day report: How do you use that information to improve your sales? The first thing that comes to my mind is perhaps regulating CPC advertising by the target hours (which is obviously valuable). Aside from that however, are there other ways to leverage the data in that particular report?nnI appreciate the 5 reports you listed at the bottom of the article; these types of reports are actionable, and can provide meaningful datannJasonn (follow me on Twitter for more entrepreneurial advice)

    • Avatar Mike Yanke says

      Hi Jason – nnIn regards to conversions by time of day, regulating the time when paid search campaigns are running would likely be the most impactful action generated from this type of report.nnIt may also be beneficial for some companies to engage in conversations via social networks during these timeframes,as well.nnHope this helps,nMike

  2. You’re exactly right, Mike. Analytics is not a perfect science, in fact it’s always changing. The bottom line is that measuring results is the most important thing. If you don’t measure it you can’t manage it. That’s the key in all successful marketing campaigns – make sure you’re paying attention to where your traffic is coming from and what ads, blogs, comments, banners, social media networks, etc are working the best. Of course, this is a moving target and something you must pay attention to continually. As you’re measuring it’s becomes as simple as “do more of what works and less of what doesn’t”. If you don’t know you’ll never grow. nBrett Relander n [email protected]

  3. I think that metrics are important, but in the end, if you aren’t getting the results you are looking for but still have good metrics, things need to change. nInformative post… thanks for sharing what you picked up at SES.

  4. Hi,

    Great insight into the world of analytics.

    What I watch the most is the referring sites, since I have a squeeze page, and Bounce rate are bound to be high. I do not consider the bounce rate to be important to my site, since it’s only page.

    Now, I do have some confusion regarding bounce rate. What I understand is that bounce rate is the percentage of people who left the site from the home page i.e. they did not visit any other webpage.

    But I have also read on a blog that it means that how many visitors leave really quickly. If visitors leave within 10 seconds, whether it is from a home page or any other page, the bounce rate will be high.

    Now, which one of these is the correct definition of bounce rate?

    • Avatar Mike Yanke says

      A bounce is generated from any visitor who enters your website, either to the home page or any internal page, who exits without visiting another page.nnThe great Avinash Kaushik offers this succinct definition of a bounce in his book Web Analytics 2.0: “I came, I puked, I left.”nnHope this helps clear things up.nnMike

  5. Website analytics are almost always tricky, especially when you have a physical location, or you are offering more than just items for sale. The tough part is to fold in off site metrics (store visits, store sales, donations, whatever they may be) to try and get a complete picture of what the web traffic and website are accomplishing. As always, measuring is a key part of marketing.

  6. Thanks for the post, very informing. I have a question though: In using Analytics I always find that there is a difference in traffic measurements from Analytics and the number of clicks I get from a Google Adwords campaign. Any ideas why there is such a difference? Just feels like Google is ripping me off sometimes.

  7. Aside from allocating budgets in paid search engines, what value do you think knowing the time of day customers make a purchase in has? How would you use that data for something other than upping your budget during those hours? Thanks for being at SES and live-blogging on behalf of those who couldn’t be there.