Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors.

bright purple Do you have a sneaking, yet unshakable, suspicion that your Web Analtyics Vendor is sometimes just trying to mess with you?

Guess what?

It's true!

All web analytics tools have a smattering of metrics and key performance indicators that were created just because someone decided it would be cute to add / subtract / multiply / divide some numbers.

Many of these don't pass the first sniff test and when if they do you are still left wondering: "What in God's name and all that is holy in this world am I supposed to action based on this metric?"

The answer?

Nothing.

With that gloriously upbeat set up let me tell you what we are going to cover today: Three metrics that are available in pretty much all "adult" web analytics tools. Daily, Weekly, Monthly Unique Visitors.

daily weekly monthly unique visitors They are so common yet most people don't understand them well enough and fewer still realize how harmful these can be to your health even in day to day use.

So in this post we try to understand the most basic of the web analtyics basics, the Unique Visitor computation.

What's a Unique Visitor?

It is simple really. . . .

Technical Definition: Count of all the Unique cookie_id’s during a given time period.

English Definition: The first time someone visits your site a first party persistent cookie is set in their browser. This cookie lasts any where from several months to several years. Each time that person visits your site that cookie identifies them as the same browser.

Notice I said browser, not person. It is likely, but not always true, that each a unique visitor is a unique person.

You can learn a lot more about Visits and Unique Visitors in this post: Standard Metrics Revisited: #1: Visitors.

Very predictably every 18 months or so the blogosphere goes wild with how accurate, or not, the Unique Visitor metric is. Much mud is thrown around. Indignations are foisted on the world. Name calling ensues.

Regardless of that Unique Visitors remains a valuable metric that used correctly, in place of Visits, measures success of your online marketing efforts.

Oh and your best weapon against ignorance? Education. See above post on Visitors. And this one: A Primer On Web Analytics Visitor Tracking Cookies. It covers cookies and deletion rates and other such yummy stuff. Read that and you have my word you'll be the smartest cookie in the room.

See what I did there? :)

Daily, Weekly, Monthly Unique Visitors:

In many web analytics tools (say Yahoo! Web Analytics, Omniture, WebTrends etc, but not in Google Analytics ) you'll also see Daily Unique Visitors, Weekly Unique Visitors, Monthly Unique Visitors and, sometimes, Absolute Unique Visitors.

monthly trend of daily unique visitors

Each is trying to tell you something about Unique Visitors, yet if you pause and think about it, I mean really pause and think about it, you'll realize two of these are really bad for your health, and the third should be used with caution.

The core reason is that what looks attractive initially becomes progressively worse as you extend the time period. The Daily metric, so to speak, does not even last in value beyond two days!

So let's spend a second understanding this slightly yucky phenomenon.

Here's the data, from omniture.com, where WebTrends is used for tracking Visitors. . .

visits by unique visitors

Now let's go measure the complex set of metrics that'll stare at you, let's say when you crack open Omniture or WebTrends (or pretty much any other competitive web analtyics tool).

The Web Analytics Unique Visitors Story:

Before that realize that what you see will depend on the time period you are looking at. [Arrrh!]

And before I really really jump in… you'll see a metric called Absolute Unique Visitor. I am going to use that as a proxy for how unique visitors should be computed correctly, regardless of what time period you are computing it for. Keep an eye on that number.

Looking at Month 1 and Week 1 at the end of Day One:

daily unique visitors

If you ran your reports at the end of day one here is what your analytics tool will report to you, with some delight and joy I might add. . .

Daily Unique Visitors: 3
Weekly Unique Visitors: 3
Monthly Unique Visitors: 3

Makes sense right? Do a happy dance, high five someone next to you, heck give them a hug and a kiss.

Now let's make this more "complicated".

Looking at Month 1 and Week 1 at the end of Day Two:

unique visitors for two days of a week

If you ran your reports at the end of day two here is what you'll see. . .

Daily Unique Visitors: 5
Weekly Unique Visitors: 3
Monthly Unique Visitors: 3
Absolute Unique Visitors: 3

Slow down the happy dance a bit.

Note the silly effect on Daily Unique Visitors, even though it was the exact same folks, Dennis and Matt, from the earlier day who visited on day two. They get counted twice.

Life lesson: Daily Unique Visitors is a useless number if you are looking at a time period of more than one day!

Let's keep going.

Looking at Month 1 at the end of Week One:

unique visitors at the end of week one

Crack open your analytics tool, it has been a long week, look at the metrics, here's what you'll see. . .

Daily Unique Visitors: 6 (!)
Weekly Unique Visitors: 3
Monthly Unique Visitors: 3
Absolute Unique Visitors: 3

Note the continuing uselessness of the Daily Unique Visitor number (and even if you trend it over time, as in the blue graph above, analyze what it is actually showing you? what's the insight?).

In your Web Analytics Tool you might see a report that looks like this:


summing daily unique visitors-no!

By know you know why there is a sad frowny face in that last Total row. Right?

Repeat: Life lesson: Daily Unique Visitors is a useless number if you are looking at a time period of more than one day!

Looking at Month 1 at the end of Week Two:

weekly unique visitors

Gather everyone in your close proximity in the office, form a circle, hold hands, close your eyes, say a quite prayer, now open your analytics tool. . .

Daily Unique Visitors: 10 (!!)
Weekly Unique Visitors: 6 (!)
Monthly Unique Visitors: 5
Absolute Unique Visitors: 5

The Weekly number is wrong because it counts: Avinash, Dennis, Matt, Matt again, Ian and Jim. It counts Matt again because he visited during both weekly time periods.

Life lesson: Weekly Unique Visitors metric is useless if you are looking across multiple weeks. We've covered above why Daily Unique Visitors is, to put it mildly, sub optimal.

Ok only two more scenarios left, hang in there, it gets better.

Looking at the end of Month 1, for the whole month:

monthly unique visitors

By now I am sure you are 100% up to speed on what you are going to see. . .

Daily Unique Visitors: 13 (!!!)
Weekly Unique Visitors: 9 (!)
Monthly Unique Visitors: 6
Absolute Unique Visitors: 6

There is now triple or double counting happening in both the Daily Unique Visitors and Weekly Unique Visitors numbers.

Life lesson: Both Daily Unique Visitors and Weekly Unique Visitors numbers are useless when you look at a time period of a month.

One last scenario, not to make your brain hurt but rather to ensure you reach the state of maximum Analysis Ninja enlightenment!

Looking at the end of Month 2, for the two months:

visits by unique visitors

Tingling with excitement. . . here's what you'll see. . .

Daily Unique Visitors: 19 (kill me now!)
Weekly Unique Visitors: 15 (can't breathe!)
Monthly Unique Visitors: 12 (!)
Absolute Unique Visitors: 9

There is now triple or double counting happening everywhere, the Daily Unique Visitors, Weekly Unique Visitors and Monthly Unique Visitors numbers.

The correct measure of unique is the Absolute Unique Visitors metric because it de-dupes the unique visitors across the entire time period you are reporting on.

Life lesson: Both Daily Unique Visitors and Weekly Unique Visitors numbers are totally really useless when you look across months. Use Monthly Unique Visitors with caution, knowing it is only de-duping for each month and then summing the number for each month.

absolute unique visitors

If your tool provides Absolute Unique Visitors you are in luck because then you are getting true unique visitors across whatever arbitrary time period you choose.

Google Analytics provides you with the Absolute Unique Visitors metric.

[Update: In the new version of Google Analytics this metric is called Unique Visitors. Everything about it described is the same in terms of it providing a unique de-duped count, it's just called Unique Visitors. You'll find it in standard reports, and you can easily add it to any custom report. Try it!]

google analtyics true unique visitors across time periods

It will do that across set time periods, like the month of March (or any number of months). . .

march unique visitors

or across arbitrary time periods, as Monday March 9th through Thursday March 19th. . .

random date range unique visitors

It will dedupe the numbers when it reports to you, rather than adding the totals of each day, week or month.

Complex but bonus for Ninjas: Depending on which graph you look at, daily, weekly or monthly, it will intelligently compute the number for each time period and also show you the aggregate deduped number for that time period.

Fly in the otherwise rather healing ointment?

Google Analtyics does not compute Absolute Unique Visitors when you segment the data, when you use the Advanced Segmentation feature. Those of you who read the blog know my utter infatuation with segmentation, so you can easily understand how sad this makes me.

You can get Absolute Unique Visitors for segments by using the "create a filtered profile that just data for the segment" method and that works if you have forethought. But it is sub optimal, just like some "enterprise" web analytics vendors telling you that you can only segment if you tell them before the fact what you might want to segment later.

Why do Web Analytics Vendors torture you with Daily, Weekly, Monthly Unique Visitors?

I knew you were asking yourself this question!

Good on you Mate.

If these metrics are that sub optimal, why do web analytics vendors put us through this torture?

Simple: Compute power (translation: cost, for them).

It is very computationally intensive to calculate for you the true real (Absolute) Unique Visitor number across any arbitrary time period or across multiple weeks or months.

Increased computational intensity for the vendor means more processing time and higher costs.

So doing Daily, Weekly and Monthly counts (and then summing them up) is cheaper for them.

After the first vendor decided to do this, and there were no major outcry from Web Analytics Users (or even Ninjas!), others quickly followed.

For the more prevalent vendors in the space Google Analytics is one the rarest that provides the truly de-duped Absolute Unique Visitor metric (in aggregate, not segmented, boo!). Only time will tell when Google will buckle under the computation/cost weight and stop providing it true Absolute Unique Visitors.

[Update: Both NedStat and Xiti, two wonderful European companies do allow for computation of Absolute Unique Visitors out of their standard packages, no additional payment or gyrations required. Add Unica's NetInsight to that list as well! Hurray!!]

There are some vendors that will tell you that you can buy their more expensive data warehouse solutions (at an additional cost on top of what you pay today) and then compute Absolute Unique Visitors yourself. True. Ask for the cost. Ask if its really Absolute. If prudent, pay more. Regardless, be informed.

Long lesson.

But now you are truly at a Analysis Ninja black belt level of proficiency!

Now your turn.

Please share your comments / feedback / critique / hugs / non-hugs about this post. What does your tool do? How do you think we should improve things? What would you eliminate? What would you add? What did I miss?

PS:
Couple other related posts you might find interesting:

Comments

  1. 1

    Well thank you Avinash for making a big note of the fact that you cannot segment AUV in GA. I really hope they find a way. I experience the same "booo" with Funnel Visualization – would be awesome to be able to apply an Advanced Segment there.

  2. 2
    Richard Vaughan says

    The company I work for uses Sitestat by Nedstat and it does a pretty good job of aggregating uniques properly not matter what the report period.

  3. 3

    I'm surprised you say that Google will buckle under the computation/cost weight and stop providing true Absolute Unique Visitors.

    Perhaps it's because I'm not familiar with what it takes to provide this information, but I would think it would go the other way. Processing power is becoming cheaper over time. Also I would think that any web analytics company worth their salt who is charging (an arm and a leg in many cases) wouldn't want a free tool like GA to upstage them.

    + + + + +
    NOTE: 1) I think you are exponentially hugely massively underestimating the amount of data GA needs to process each day. 2) Please know that I am just saying if every other vendor gave up on providing this metric, Absolute Unique Visitors, in their standard web analytics tool (Site Catalyst, WebTrends, Yahoo! Web Analytics etc) then maybe Google will follow. BUT. I am not saying it is happening or it will happen or should happen. : ) -Avinash.

  4. 4

    Thanks Avinash for a very detailed post on Unique visitor.
    This is indeed another great example to show that we need to consider very carefully when trying to interpret Analytics data.

    There is another similar concept that we use internally "visit" vs "landings" (landing = initialization of a new traffic source) which create similar problems when looking at different time frames.

    From this part of the wood (SouthEast Asia region), only a very small portion of the advertisers is at this level. Most of them, unfortunately, use Web Analytics to a very limited extend.

  5. 5

    many many thanks for this post. I'll bookmark it to give my clients a powerful example on a thing I'am not able to explain as well as you :)

  6. 6
    Ned Kumar says

    Agree with you Avinash on the uselessness and dangers of aggregating Daily Uniques. And on trending, I would say that if someone really really feel the urge to trend DUVs :-) then do it for the incrementals – so in the example above, 3 DUV the first day, 2 DUV the second day, 1 DUV the third day and so on.

    Also, on the case of unique visitors in general – if one is lucky enough to have a userid/sign-on/login etc. information for their site (even for a subset of their visitors), and they have the infrastructure where they are passing that into their WA tool – then I think they should leverage that and use that in synthesis with and complementing the cookie based counts.

    Of course, there is always the KISS principle (Keep It Simple Stupid). I agree with you totally that just using the WA tool (and GA to get a leg up :-) ) and using the cookie based counts wisely can still provide one with useful information — so long as they understand all the caveats and assumptions involved. Or on the other extreme one can do some fancy number crunching and analytics using information from the WA tool and something like SAS or Tableau.

    (I would like to cut down on the categories though :-) (from DUV, WUV, MUV, AUV etc. to just UV).)

  7. 7

    Avinash, I am so glad that I subscribed to your RSS feed a while back. I truly enjoy this type of "back to basics" type of post. A fantastic reminder to use actionable data and ignore the fluff.

  8. 8

    Hi Avinash:

    As well as you we are too pretty sad about the AUV's non availability when using advanced segments.

    The problem is that the work-around by using filtered profiles, don´t work well when you are firing events, because the events are not easily related to visits and is required too much computation power as well.

    As you mentioned into the Ninja's Bonus when you look at the Visitor Trending -> Absolute Unique Visitors report into GA you have booth metrics: the Absolute Unique Visitors number for the whole period and the (un-duplicated) unique visitors number for every single day/week/month. Many people don't realize that if you add the unique visitors number across the whole period it differs from the AUV showed into report.

    By the way, thanks for your conference into Buenos Aires, was such an inspiration for all us.

    Regards,

  9. 9

    Hi Avinash,

    Great post that certainly disentangles the multiple-unique-visitor-metric dilemma.

    Now that you mention the frequent discussion on Unique Visitors accuracy on the blogosphere, which is your take on the UV inflation?

    1) How much of a problem do you think it really is?

    2) How about the impact on visitor loyalty metrics?

    3) Are UV based metrics useful only to show you the trend but not the magnitude?

    4) Should we better be using Visit based metrics? Or perhaps use other correlations and paralel metrics to validate the trends and ranges of magnitude?

    Thanks,

    Jose

  10. 10

    Hi Avinash- great post!

    It saddens me to hear that you wonder how long it will be before GA takes away absolute unique visitors!

    I've also had many "awwww"s as I've tried to get absolute unique visitors with segmentation to no avail. However, I always had faith that it would come someday, as GA seemed to be improving with each new release.

    So with this post, the Web Evangelist himself, saying it probably will get worse, I'm truly saddened. Say it ain't so!

    Thanks,

    Nelson

  11. 11
    Simon Tu says

    It is amazing that you can make the complicated so simple to understand Avinash. Wonderful post.

    You were quite restrained though. For example our famous enterprise vendor only provides Daily Unique Visitors by default. Before I joined our company was reporting totals for daily uniques by Week and Month, which as you point out is complete garbage.

    We had to pointedly ask for Weekly and Monthly Uniques and were told it could be "turned on" for us but would cost more in terms of the contract.

    Many vendors continue this deceptive practice even now, after daily uniques have been debunked so thoroughly.

    Simon.

  12. 12

    Another great post, Avinash!

    Very thought-provoking. This is an issue I go back and forth on and I'm constantly struggling with it.

    Here are my issues:

    1. Cookie deletion is only half the problem – different computers/browsers/devices are a big problem, too.

    2. Going along with #1, although *we* understand the complexity of unique visitors, the people who demand these numbers probably don't (and explaining it gets you glassy-eyed stares!). My problem is with the name – unique visitors – which assumes we are talking about unique people.

    Anyway, some thoughts.

    Shelby

  13. 13

    Couldn’t agree more!

    AND since my good name is used in the example above, I feel almost obligated to respond :-)

    Daily, Weekly and Monthly unique visitors was a great way of faking it for a while (as it is indeed more “expensive” to provide true visitor level segmentation). I was quite honestly impressed by GA launching with this way back (actually cried a bit on the inside as we were still a independent paid for vendor).

    But I am happy to announce that upcoming version of Yahoo! Web Analytics, will indeed hold an amalgamation of the above three into ONE “Visitors” metrics.

    So for comparison, and to avoid confusion:

    Google Analytics — Absolute Unique Visitors
    Yahoo! Web Analytics — Visitors

    The super cool thing about this release, is that we will compute Visitors (or in GA terms Absolute Unique Visitors) when you segment, filter or otherwise manipulate your data.

    But much more about this release shortly!

    Cheers
    d. :-)

    Dennis R. Mortensen, Director of Data Insights at Yahoo!
    Blog: http://visualrevenue.com/blog
    Book: http://visualrevenue.com/blog/yahoo-analytics-book
    Twitter: http://twitter.com/DennisMortensen

  14. 14
    Jlbraaten says

    Avinash,
    Another great post. It is unfortunate that the folks holding the purse strings in many organizations aren't aware of these subtle nuances between analytics vendors.

    Often, a price tag is indicative of quality in their eyes… never mind the 90/10 rule.

    Thanks for continuing to arm me with some of the tools necessary to wage my Evangelism (and Ninja) campaign.

    -Josh Braaten

  15. 15

    Andrew: I think you are exponentially hugely massively underestimating the amount of data GA needs to process each day.

    [+Nelson ] Please know that I am just saying if every other vendor gave up on providing this metric, Absolute Unique Visitors, in their standard web analytics tool (Site Catalyst, WebTrends, Yahoo! Web Analytics etc) then maybe Google will follow. BUT. I am not saying it is happening or it will happen or should happen. : )

    Chandler: That's a nice concept, visit vs landings. Understand that helps you focus a lot better.

    Ned: Are you sure you are not putting lipstick on a pig by doing incremental trending? :)

    Just teasing you!

    There are a number of ways to ensure you are capturing the best possible dataset. Using first party cookies. Doing correlations with your login data. Even once in a while mashing it up and having a fight with your log file data.

    Let's follow all the best practices, the data is meaningful and it helps make good decisions. No?

    Jose: I do not want to shortchange my answer to your question about deletion rates etc, please see this post:

    A Primer On Web Analytics Visitor Tracking Cookies

    It answers your questions.

    Simon: I did not want to be impolite, hence skipped that part.

    Shelby: I am sympathetic to the fact that the metric's name could be cuter.

    That does not distract from the value of the underlying measurement when you collect the data using best practices (see reply to Ned above).

    No metric in this world is a panacea.

    Visits (sessions) is a wonderfully useful metric, it is everywhere and should be used to the fullest extent of the law!

    Ignoring Unique Visitors means making a decision that pan-session behavior is not important and that is a very sub optimal strategy by anyone. Or that measuring Conversion Rates as "one night stands" is the way to do marketing. Sub-optimal (I am sure you would agree).

    Visits can't replace Visitors, and vice a versa. Each holds a valuable insights, it is your job to identify what that is for your business (edu) and it is mine to do the same for my business.

    Of course you know all this already, I am just restating the obvious. : )

    Dennis: Hurray! Can't wait.

    And now you have shown that a free "enterprise" tool will provide data / metrics that even paid "enterprise" tools can't match (yes yes they can match it if you spend an additional $$$,$$$.$$ on top of what you are already paying). Sorry could not resist.

    I am honestly thrilled.

    -Avinash.

  16. 16

    Hi,

    Nice post. Something similar puzzles me.

    Let's say I have excluded keywords in the non paid keywords report like kw1|kw2|kw3 via the box at the bottom.

    When I look at the data monthly for a quarter it is like jan 1000, feb 1200, apr 1300. But when I look at the quarter in one go even though it should give me 3500 total, it doesn't add up to that – it is less.

    I understand and know the logic behind UV counts as stated in the post but I couldn't pin point the reason for this.

  17. 17

    Great explanation of the problem!

    GA calculates the Visitors right, but as long as users copy the numbers to their Excel sheet once a month and sum them up there, it won't help :-(

    Maybe the vendors could introduce some marker for "unsumable numbers". A little icon or color. Just a thought (maybe for Dennis *g*)

    -Markus

  18. 18

    Great idea to clarify daily, weekly and monthly unique visitors calculation.

    AT Internet/ XiTi customers take benefit of this feature for many years now.
    Whatever the time period you select in the calendar, our solution gives unique visitors number.

    Yes you're right, an additional calculation is necessary, it takes a little time on long time period but you get data in Behaviour > Unique visitors > Site summary. And you can apply segmentation on this report…

  19. 19

    Great discussion. Just want to add one more parameter: In a western country with very high computer density, like here in scandinavia, a great part of the users have at least 2 computers, at home, at work, and maybe a laptop, an Iphone, and a minipc… and, like me, have at least 3 different browsers in each PC … how many absolute unique users am I?

    So we have removed focus from uniquness long time ago. A visit is still the most important business metric. By the way, when we can, we track logged in memberIDs and calculate a uniqueness for other traffic based on this.

  20. 20

    Hello Avinash,

    It is indeed a great pleasure to see a post about unique visitors from you.

    I know I am quibbling with a Guru but in some cases counting daily unique visitors in a month's trended data is not wrong, directionally speaking.

    The following assumptions have to hold though.

    1. The underlying data stays on a similar pattern of crests and troughs over time. Everyone will have noticed the sine wave that comes across one's weekly data where visits peak on one day every week.

    2. The rate of return of visitors does not change over that time frame. So, the proportion of visits/visitors who come to the site daily, weekly and monthly does not change markedly over time.

    Again, thanks so much for keeping up this blog. It is appreciated so much!!

    Regards
    ateeq

  21. 21

    Hi Avinash,

    This comment isn't totally related to this post (sorry!) but after searching Google for 1hr to no avail, I thought you might have the answer…

    Do I count 'flashservices/gateway' (flash remoting) as a page visit? I think it's just a way of connecting to flash content already displayed on a page.. but am not totally sure. Do I count this metric or remove it?

    Thank you!
    Mary

    PS. I'm reading your book for a uni assignment and it is great! :)

  22. 22
    Scott Zakrajsek says

    This is a great post on a question that seems to surface over and over again. I agree with Ned that there is benefit in tracking unique visitor counts for incremental time periods, daily, weekly, etc (despite putting lipstick on a pig). However, one problem we run into is that our WA data is rolled up based on a 4-5-4 fiscal calendar (this was our choice to appease the finance gods, everything must map to the fiscal calendar). In this situation a 5-week month could falsely show a huge traffic boost over a 4-week month, etc. Just another thing we have to remember to educate folks on. Thanks again for another solid post.

  23. 23

    “What in God’s name and all that is holy in this world am I supposed to action based on this metric?”

    Hmmm. Wrong question. :-)

    The question is typically phrased by senior management as (for info sites vs selling sites) "How many people visited?". The sometimes implied sub-question being "how many left happy and satisfied with our offerings".

    Which in turn means that if a simple metric like Visitors, in whatever variation, has a downward trend, then action that *will* be applied is: This site isn't working anymore, shut it down; and you're now out of a job.

    So, yes I agree from a tactical perspective, these broad brush metrics aren't useful. From a strategic perspective they're often the only ones that matter.
    These metrics are a … simplistic proxy for the detail of page landings, bounce, seo etc etc etc etc. Treat them as such, and they are incredibly useful; powerful even. Ignore them – imho :-) – at your peril.

    Cheers!
    – Steve

  24. 24

    Hi Avinash,

    What about the daily trend of monthly or weekly unique visitors. In other words, plotting the monthly or weekly unique visitors for every day of the month or week.

    Media sites I work for use it often, since it allows them to see how many new users they capture every day, in the context of the month (or the week).

    In other words, it sort of answers the question: do we capture new users, or we are stuck with the same audience over and over again?

    Looking at the daily trend of daily uniques can sort of do it, but I rather do it with the monthly/weekly metric, since visitors do not necessarily come back every day.

    What do you think about it?

    In my opinion, it's the only advantage I see about having the weekly and monthly pre-calculations as standard metrics you can trend daily. This is something I miss with Google Analytics (even though it's not at all a priority inmho, the use is very specific)

    Cheers
    Pere

  25. 25

    Avinash,

    It is amazing that even in 2009, we could still learn a lot by zooming into individual metrics (unique visitors, conversion, pages per visit, avg time on site, etc.) and have a nice overview. I wish I had this article when I had to explain this concept using HBX when internal client wanted quarterly visitors…

    Btw, thanks for sharing an article (including your old one) after I reminded you of a bitter article. I guess the bottom line is, there are so much learnings and reminders from topics including unique visitors to conversions (outcome per unique). Again, thanks for this great article on unique visitors and a heads up through your email.

  26. 26

    Nit: The total in your example won't tie because it is possible, just to be a bit extreme, that the same 1000 people from Jan came again in Feb and Apr. In this case your Absolute Unique Visitor number would be 1500 (1000 from Jan, 200 from Feb and 300 from April).

    Benoit: Thanks for adding, I totally forgot!

    I have updated my blog post to clearly indicate that Xiti has this functionality (as does NedStat).

    Ateeq: I can agree that if the assumptions you mention hold true then that might be value in trending the data.

    My hope though is that the Web Analytics vendors will solve the core problem and allow all of us to use the right data better rather than having to make accommodations for sub optimal data that is shoved at us.

    Scott: If other solutions did what Google Analytics, NedStat and Xiti do, allow you to create your own Absolute Unique Visitors across any arbitrary time periods, then your problem will go away.

    It is certainly the right way to do things and my hope is more vendors will do that.

    Steve: I am confused, but let me give it a try. . .

    There is no People in Web Analytics. (Is that even grammatically correct?)

    I would answer that question by giving a one min elevator pitch on what UV's are and what they measure. Unlike many I continue to believe in the value of using Unique Visitors (they are not the panacea, but they are not the dirty diaper as others have said).

    Next if indeed the action taking is as extreme as the one you have suggested, based just on Visitors, then it might be time to look for another job because that kind of management will run any business to the ground? :)

    Pere: If there are specific campaigns I am running to acquire new traffic (say I am going to plunk down $1 million to spend on FaceBook in the next 9 days) then I think there is value in looking at the daily unique visitor number trended over those 9 days (or those 9 days and the next 18 to catch the laggards).

    But unless a stimuli like that exists the daily trend will always be a poor answer to the question: "are we capturing new visitors to our site".

    As you correctly mention weekly is better, monthly is better still.

    Not just because the data will be better but also because it will force the Ninja to do strategic analysis and not be a Squirrel and chase his / her tail by doing tactical analysis on daily data.

    -Avinash.

  27. 27

    Dear Anivash,

    As part of our reporting for B2B web analytics – the most difficult aspect is to present the website visits data in the right context. The numbers really make no sense to the business and core marketing community. Yesterday I came up with this simple concept and thought to have it weighed by you ( just in case you happen to see this in the million chatter). Probably it is nothing new!
    The website visits can we presented in 4 main clusters

    1) Brand Recall Reorganization: Number of direct page visits can be categorized in this bucket. So any variation in these numbers would mean increase or decrease in brand recall.

    2) Search Engine Mindshare: The total search traffic is a measure of the competitive visibility that my brandcompany has on the internet. It’s a measure for the SEO team and also for content creators. The index can be used by all teams to see how they are faring MoM on search engine prominence.

    3) Brand Awareness: All website visits associated with campaigns can come under this tab. Marketing campaigns should also focus on creating awareness measured by unique visits from campaigns whether its Google, or email traffic. Leads though remain the primary campaign objective – though its difficult considering 6 to 24 months sales cycles.

    4) Online PR: Any referral traffic could be reported as a matrix of sum total of all visits from social media visits, online press releases, columns and articles across portals.

    This is just the first brain wave. Do you think it makes sense to explore it further.

  28. 28
    Srividya says

    Amazing Post, for a newbie analyst, things can't get clearer than this. Been dabbling with web analytics tools and I am amazed how each one is similar but yet so different and how we have to be careful with our insights.

  29. 29

    Avinash,

    Thanks for posting this. You could also take a simple Venn diagram approach to explain this concept. The absolute unique visitors are:

    V(a) + V(b) – V(aUb) where V is the number of visitors, a is the first period and b is the second period. Absolute unique visitors are the visitors in period 1 plus visitors in period 2 minus the common visitors in period 1 and period 2.

    Cheers

  30. 30
    Jim Maffessanti says

    How enlightening!

    What are your thoughts on panel based web analytics like ComScore, Nielsen NetRatings?

    I feel they under-represent "at-work" usage.

  31. 31
    Sarah Madison says

    who cares about any of this stuff, its trends that you measure anyway, it doesnt really matter what they are called.

  32. 32
    Massimo Paolini says

    As an Evangelist in training your blog has been and continues to be an incredible source of education.

    This post addressed a specific question I had about absolute visitors and the disparaging numbers they showed, and the keyword here is _absolute_.

    Thanks,

  33. 33

    You state at the top that Daily Unique Visitors are "bad for your health". I'm wondering about operations whose audiences are inherently short-lived, perhaps even 1-offs, e.g. Conventions, Concert Ticket vendors, Online Fundraising drives, American Idol polls. In these instances, a 48-hr window is all that matters, as repeats are mostly non-existent.

  34. 34

    Why, oh why couldn't you have written this post a few years ago?! :-) I have a very clear memory of a frustrating, circular discussion with WebTrends back when I was still learning the ropes, because the support fellow I was talking with didn't even fully understand the calculation and storage of Uniques.

    You nailed it, in that it's a matter of aggregation tables simply incrementing daily, weekly, monthly counts as a way to shortcut data storage and processing. So, they weren't able to de-dupe to get to absolute uniques. Once that lightbulb went on, I immediately wanted to run and yank all occurrences of the Daily Unique Visitors that our end-users were able to get to, because they were happily tripping along and looking at the "Total Daily Unique Visitors"…for the month. Egad!

    As for @Sarah's comment (No. 35) — that's a little scary. On the one hand, yes, all web analysts have to understand that trend analysis trumps "perfect accuracy" in web analytics. BUT, this post is a great example where a metric can be off by an order of magnitude from "reality" and, thus, drive some pretty ugly decisions (memories of working with a client who bought a marketing automation tool largely based on a gross misunderstanding of the data they were getting from a log-based installation of Urchin with no bot/spider filtering at all…and the implementation of the tag-based marketing automation tool surfaced the mistake ).

    Great post!

  35. 35

    Apuarv: I think the framework you are thinking of might be interesting and of value.

    I am not sure about the names of the buckets. They often tend to mean things different from what you have defined them as, and that is always sub optimal.

    I am not a fan of "sexifying" name, more there: “Engagement” Is Not A Metric, It’s An Excuse

    Jim: The core things to consider for panel based measurement are:

    1) "Extrapolations & Mathematics" – the most famous online panel for the use measures online behavior of 200 million Americans using a panel of 180k. Check what the number is for your panel.

    2) Sampling Bias – the most famous online panel is banned from installing its "monitoring software" (pseudo-spyware) at pretty much all companies (work) and Universities etc, it is also intrusive in terms of data it collects so even most minorly web savvy folks won't install it.

    In general Panel based data starts to get decent for sites that get more than 2 million Unique Visitors a month.

    At the end of this post you'll see me talk more about Panel based methods:

    Competitive Intelligence Analysis: Google Trends for Websites

    And here:

    Competitive Intelligence Analysis: Google Ad Planner

    I think this article highlights how panels impact not just online but the limitations they have offline (which are much worse online for them):

    "The TV business rests upon the central lie that it knows what ur watching."
    http://twitter.com/avinashkaushik/status/1589775295

    Sarah: I love trends, they have a lot of value: Be Awesome At Comparing KPI Trends Over Time

    In this case if you follow what's happening in the tables you'll quickly realize that the trends of these data will show crap. That's my worry.

    If that can be accounted for then absolute numbers don't matter, trends do.

    Hasan: There are certainly exceptions to the rule. But even in your scenario would you not want to measure how many Absolute Unique Visitors you got over the time period you had tickets on sale or were raising funds? My personal take would be that even in those scenarios knowing Daily Unique Visitors would serve no value.

    For example let's take this egregious scenario. You announce a sale today. It is a one day sale. Tickets are only available for tomorrow. That's it.

    So you could go back and say, how many unique visitors I got on that one day? That's a good question. You can answer it by choosing Absolute Unique Visitor metric for that one day. The presence of the Daily Unique Visitor metric adds little/no value even in this scenario, and of course is corrosive for all the other scenarios in the post.

    Would you agree? No?

    Tim: You are so kind, thank you.

    -Avinash.

  36. 36

    This is an excellent explaination of the problem with our existing analytics packages.

    Business (non analytics geeks)users look at something like "Daily Unique Visitors" and assume it is an accurate representation of individual people over that period of time. That is wrong on two fronts. One for the reasons you have shown. It is also wrong because cookies do not equal people.

    There is a movement in the Web Analytics standards bodies to require any measurement of "unique visitors" to be based on some identification of real people. It is a lot more work to create that metric since you have to understand (discover) the relationship between people and cookies for your site. That is a major exercise to get that data but probably worth doing if unique visitors are that important to your business.

    The current usage of "unique visitor" is really "unique cookie". Changing the name on reports would help if it resulted in calls from people who don't understand what it means. That creates the opportunity to educate the users of our data on that distinction. Some will be able to understand the issue and appreciate the difference. Others will not want to get into that level of detail but at least they were not misled by a false label.

    This requires changing the names of the data item as it is displayed on reports. We don't always have that ability but we should be demanding it from the vendors who are accepting our money.

  37. 37

    I love a good expose! This post definately sheds new light on the most widely used metrics (and their measurements).

    How could we have gone this long without knowing the truth behind their nefarious lies!? I'm outraged. Thanks for bringing justice to the analytics world, one metric at a time.

  38. 38

    Ken: Thanks for the feedback.

    On the point of measuring people, it is simply impossible. At the moment.

    Anyone who does not accept that is not fessing up to reality.

    We can certainly correlate with panel data. We can install spyware on our users computers. We can make everyone on our sites log in. We can have a non-blowable persistent cookie. So on and so forth.

    Each method has its flaws and we'll be no closer to making decisions if we wait for perfection.

    At the moment we have a good enough method, using first party persistent cookies. That method respects the privacy of our users, it does not freak them out (ok, there are some it still freaks out :).

    This method, first party cookies, is not perfect. It will improve with time (both because of demographic and psychographic changes, and because of technology changes).

    Until such a time my vote is to make decisions every single day with the good enough because it is of value and still better than any data you get from any other channel on the planet.

    -Avinash.

  39. 39

    Avinash Thank you for this. now I understand more bout the idea of site visits

  40. 40

    Thanks Avinash for the Insight.

  41. 41

    You said, This cookie lasts any where from several months to several years.

    My cookies last until I close my browser, assuming I have them turned on in the first place.

    Why not just use page views? At least you can tell if your website is getting more popular or less.

  42. 42

    Interesting points as always, Avinash, thank you.

    What's your take on the way Webtrends uses Dynamic Visitors, a measure that switches dynamically between Daily, Weekly, Monthly, Quarterly and Yearly Unique Visitors based on the time frame selected in the report?

    Obviously there's still the potential for some double counting (particularly since any custom date range will be built by adding smaller date ranges together), so you've got to bear that in mind. But when we look at a week or a month or a year, and Webtrends provides the appropriate unique (de-duped) visitor number for that time period, that's a fairly worthwhile metric, isn't it?

  43. 43

    please enlighten me: what is the difference og google analytics ang analytics using th logs?

    Which is more preferable?

    What are there main differences on the data?

    Thank you

  44. 44

    Avinash, excellent post, as usual. Thank goodness that "Angie" person didn't get lumped into the example until the end, because she's a habitual cookie-deleter and would really mess up the numbers. Although I guess she wouldn't have messed up the DUV numbers because they're messed up anyway. ;-)

    I'd like to point out that WAA Standard definitions do NOT recognize the concept of "daily unique visitors," etc. as described in your post. Just "unique visitors," and they aren't unique visitors unless they're de-duplicated over the reporting period,regardless of what that reporting period may be.

    We can add Unica's NetInsight to your list of tools that does this calculation properly, de-duping over the requested date range.

    I'd also like to point out that, although GA uses cookies for UV tracking, some of us track authenticated userids in other tools, along with cookies. And most of the for-pay tools have some sort of hierarchy where they'll substitute some other identifier for the cookie if it's blocked (session ID, IP/User Agent, etc.) and some will use those numbers to take a guess at "anonymous" unique visitors. So to address Ken's comment above: unique visitors and unique cookies are not the same thing. That may be true for some tools, but it's definitely not a universal truth.

    Also to Ken's comment, there is *not* a movement in web analytics standards bodies to require tying back cookie information to people… at least not that I'm aware of. :) That movement is driven by the IAB, which sets *advertising* standards. They have different reasons for doing what they do, and they don't represent most of us who are doing web analytics for a living, only that part of your time when you might be wearing an advertising buyer/seller hat.

    Avinash, I agree with you 10000% about measuring "people" — I just don't see it happening any time soon, and as a practitioner I don't want to tell people they aren't welcome at my company's websites unless they tell me who they are. If you know what you're doing, you can still make some valuable observations based on cookies. No, you don't know how many people are visiting, but you can indeed get some insight into cross-session behavior and loyalty trends.

  45. 45
    Abdullah says

    Excellant post Avinash , I use most of them but I do trust google analytics only "Lucky man" :D

    what do you think about AWStats ?

    Thank you

  46. 46

    The media industry is in love with daily, weekly, and monthly unique visitors. Probably because it mirrors what's available in external measurement offerings via comScore, etc. If taken on a trended daily/weekly/monthly granularity report, those breaks actually work quite well.

    I totally agree with your "up in arms" comment about unique cookies. Who cares people! Get over it already! Use it as a tool, as a relative measure, don't use it, I don't care. There are more important things to worry about, like conversion.

    Regards,
    Garry

  47. 47

    Hi Avinash,
    Thanks for a great column. Would you please say again, slowly and loudly, how one would go about creating a filtered profile to produce AUV? It is not on the pre-defined filters list, needless to say.
    Thanks/Amy

  48. 48

    Avinash,

    Excellent post. Can you tell us what a Unique Pageview is? Is it equal to a visitor? Equal to a unique visitor?

    Thanks,

    BRLM

  49. 49

    Unique page views have nothing to do with unique visitors.

    Let's say you come to my site and have this session:

    Visit 1: Home page. About page. Videos page. Home page. Latest post. Videos page.

    That's:

    1 Session.
    6 Page Views.
    4 Unique Page Views.

    Hope this helps.

    -Avinash.

  50. 50

    I've gotten so much value from your blog. Thanks for all the good work that you do.

  51. 51

    Great post. Omniture's Discover does compute unique visitors across a given time period – but your point about providing it as part of the standard package is spot on.

  52. 52

    I completely agree that uniques are misinterpreted if you aggregate them. However, they still provide tremendously value if you use "Average Daily Uniques" or "Average Monthly uniques" across many days or months.

    For example, I could have a quarterly dashboard that shows average daily uniques during Q1, and watch how it grows compared to Q2.

  53. 53

    Rich: I am sorry but I am unsure of what strategic or tactical value exists in looking at "average daily uniques" for one month.

    As I mentioned in the post I can see how someone might report things like "during this week our Average Daily Uniques" were xyz.

    But once you move beyond a week it is not very useful (far too much double counting and other stuff) and certainly for a month the underlying data is such a mush that it is not of very much value.

    Many tools, excepting Google Analytics, by default tell you Average Monthly Uniques if you look at time period longer than a month (GA will still give you absolute unique visitors). In that sense you might not have much of a choice when you look beyond one month.

    -Avinash.

  54. 54

    Thanks Avinash for the feedback.

    I use Omniture SiteCatalyst and it has the Yearly Unique Visitors metric available. Would you then recommend that we should almost always use this as the metric to measure traffic? This is essentially the same as Absolute Unique Visitors if the reporting timeframe is captured within the same calendar year, correct?

    (Obviously, we run into duplication of visitors issue if we a report for a timeframe across multiple years.)

    Thanks!

  55. 55

    Rich: If you are reporting on a year time period or conveniently report on data from, say, Jan to April 2010 then you can run the report on May 1st (and make sure you run it on May 1st or you are out of luck) and use that number to report the count of unique visitors to your site during that time period.

    Usually that is not convenient. You don't always want to report on YTD or just the year.

    Consider these extremely common scenarios:

      You probably want the count of unique visits from Jan to March 2010 (and you want that number any time, not just on April 1st).

      Or you want the unique visitors for April & May 2010 because you ran a massive campaign and you want segmented (say Paid Search only) unique visitors.

      Or you want unique visitors from Feb 15th to April 15th because it coincided with launch of your new products.

    All of the above scenarios make the limited Monthly or Yearly unique visitor numbers from your analytics tool to not be of much value.

    All of the above scenarios you can do with Google Analytics. You can get Unique Visitors for any arbitrary time period and you can get Unique Visitors for arbitrary time periods for any segment you want (simply create a custom report and apply any segment you want on the fly).

    That is what you want.

    Not to adapt to the tool rather adapt the tool to your business needs.

    Please note that at an additional cost you can purchase more tools from Omniture like Discover 2 or Insight in addition to Site Catalyst then you can create a query that will allow you to do the above. You just can't do that with Site Catalyst.

    -Avinash.

  56. 56

    I think I'm missing something and hope you can clarify: in your first paragraph directly above this post: why do I have to run the report on May 1? If I visited the site on Jan 3rd, and again on Apr 5th, and then again on May 6th……and say today is May 15th, why can't I still run the reporting timeframe of Jan-Apr 2010 today (on May 15) and still get accurate numbers? I'm still counted as 1 yearly unique visitor (during Jan-Apr 2010) which is equivalent to Absolute Unique Visitors. I don't see the importance of why must I run it on May 1, 2010?

    The other examples do make sense though: say I wanted unique visitors from Feb 15 to Apr 15 (as in your example above) – I would NOT be counted as a unique visitor in this bucket for the Yearly Unique Visitors report if I came to the site prior to Feb 15 during the year since I would have been counted as a unique visitor prior to Feb 15. This unique visitor count will then be "under-reported" since the yearly unique visitor count records the "uniqueness" upon my first visit to the site during the year.

    Thanks again for the help! Appreciate you being a thought leader and educator in the world of web analytics!

  57. 57
    Marco Remmerswaal says

    Hi Avinash,

    I have a little question. Isn't it true that you can't really measure unique visitors because of the following reasons:

    – a lot of people/companies clean their cookies frequently. When you come back to the site the web analytics tool thinks that you are a unique visitor.

    – you can use login accounts to measure unique visitors, but one or more people can you the accounts, so it's not 100% sure.

    – One or more people can visit a website from one computer. These are all unique visitors but not measured like a unique visitor.

    Isn't it better to don't report about unique visitors at all?

    I am very interested how you think about this.

  58. 58

    Marco: The statements you have expressed are overtly broad, even as they express some parts of the reality we all have to live with.

    Unique Visitors can be tracked, and to an extent that the number is quite useful for most websites. On to your questions….

    "a lot of people/companies clean their cookies frequently. When you come back to the site the web analytics tool thinks that you are a unique visitor."

    It depends on the type of the website, but frequently is perhaps stretching things a bit too much. For some tech savvy sites, say slashdot, that have a unique type of audience the cookie deletion can be up to 30% or even more. But for most websites that is not the case, for first party cookies (if you are still using third-party cookies then may the lord save you! :)). For this blog cookie deletion is around 3%.

    Please read this blog post for indepth information about cookies. It is important that you get educated: A Primer On Web Analytics Visitor Tracking Cookies

    "you can use login accounts to measure unique visitors, but one or more people can you the accounts, so it's not 100% sure."

    Yes you can, and you should use features like Analytics Custom Variables to track logged in and not logged in behavior.

    But remember that even when login is mandatory (like with banks, not like amazon where it is not mandatory), a tiny percent of your site traffic will log in.

    Which means that can't be your only method of collecting Unique Visitor data since you'll only be analyzing a tiny fraction of your website Visitors.

    "One or more people can visit a website from one computer. These are all unique visitors but not measured like a unique visitor."

    Yes this can happen, especially in countries like India and China where availability of computers is not as much as it should be. In many of these cases the browsers also flush all the cookies after each session (for the next person who will log in), which also is something to consider.

    So depending on the country you have to be a bit more careful. This for example is much less of an issue in the US or EU. Either way if you want to put in the effort you can analyze your logs and your tag data and run analysis to get a feel for how much this issue bedevils you.

    "Isn't it better to don't report about unique visitors at all?"

    Perhaps in some cases.

    But in most cases the data you collect is sufficient and reflects enough reality that it will be of value and you should use it.

    We have far too much "quicke" mentality in Online Marketing. We simply shout "Convert!!", "Convert NOW!!!", "WHAT THE HELL ARE YOU WAITING FOR THIS IS YOUR FIRST VISIT WHY THE HECK DON'T YOU CONVERT ALREADY".

    It is painful to read that I know. But that is how we do marketing today.

    When we focus on optimizing our sites through analysis of Visits that is exactly the kind of mental model we are empowering. We are optimizing for Visits. The reality is short of buying Pez Dispensers people don't convert on the same visit, they might never convert.

    I am a fan of being customer centric. Reporting Unique Visitors and using Unique Visitors in metrics like Conversion Rates (Orders/Unique Visitors) helps us focus on multi visit behavior. I like that.

    As with all metrics in Web Analytics, Visits or Page Views or Time on Site, accuracy is relative and focusing on trends is far more impactful and the quest for 100% perfection is futile.

    -Avinash.
    PS: One final way to check if measuring and analyzing Unique Visitors is right for your business…. Measure Visits to Conversion (/Purchase) reports in your web analytics tool. If you are really top heavy then you might get away with it, but if not then your Visits only focused analysis will yield undesirable results for your company.

    Here's a helpful post: Excellent Analytics Tip#6: Measure Days & Visits to Purchase

    • 59
      Robert Bruce says

      What happened to Absolute Unique Visitors? I'm not seeing it in the new version of Google Analytics.

      Did Google cave here or is Unique Visitors "smart" now such that it de-dupes in the V5 version?

      • 60

        Robert: With the newest update in Google Analytics the word Absolute has disappeared and the metric is now called Unique Visitors.

        But it is Absolute Unique Visitors, deduped across any time period (fixed or arbitrary). New name. Same goodness. : )

        Avinash.

  59. 61
    Joao Torres says

    I'm totally agree about this kind of analysis. But I think that we need to understand web as media, not just auctions, e-commerce and these things. The digital market wants to increase the advertisers investment. Unfortunately, the web is not e-commerce overall and how can we build safe measurements tools to understand what the advertisers needs, because brands can be developed on the web, a huge part without on line purchase just to communicate. If consumers will decide to buy this product next weekend on the real shopping center, should we not regard it?

  60. 62

    Joao: The web can indeed be used for multiple purposes, to drive multiple outcomes. That is part of its awesome glory.

    Here are two blog posts on measuring multi-channel impact using various analytics techniques:

    Online impact of offline campaigns: http://bit.ly/akoffon
    Offline impact of online campaigns: http://bit.ly/akonoff

    And here is a post about how to do web analytics on a Government website (as you know no ecommerce or even leads or anything):

    Web Analytics Success Measurement For Government Websites
    http://zqi.me/akgovt

    Finally here is a blog post that might be particularly apt in your case, brand measurement (as you indicate "communicate and inform"):

    Brand Measurement: Analytics & Metrics for Branding Campaigns
    http://bit.ly/akbrand

    Hope this helps.

    Avinash.

  61. 63

    After reading this overview I think I have a better understanding of how Unique visitors number is calculated. But what is the difference now in GA between "Visitors" and "Unique Visitors". I'm doing some custom reports for our sales team (they want to know monthly unique visitors to the site…), and I keep getting the Unique Visitors number that is higher than Visitors (holds true across all segments, including unsegmented "all visits"). Is there a glitch in GA? Which number is more reliable?

  62. 64
    karishma says

    Hi Avinash

    Thanks for this great post. However my current need is to understand Daily Returning User Vs Monthly Returning User over a period of time say a month. how to de-duplicate these metrics?

    This will really help me

    thanks

  63. 65

    Hi Avinash,

    i have a question concerning the two metrics "visits" and "unique visitors" in google analytics (no filters, no segmentation). Most of the pages of our website have slightly more Visits than unique visitors. I think that is normal. But some pages have up 10 times more unique visitors than visits. How is that possible ? What are possible explanations? A problem with cookies or analytics code?

    I found the same problem on varios blogs/boards but no answer…

    Thank You very much for an idea!
    Peter

    • 66

      Peter: I'm assuming you are looking at the Page reports in GA and then looking at Visits and UV.

      You would think that UV can never be greater than Visits, and that is true. But there are quirks. In GA Visits are counted on the first hit of a session. So if a page is not viewed as the first hit of a session it will increment the count of Visitors but not Visits. Hence, in some cases, you end up with Unique Visitors > Visits.

      Optimally what you want to measure on your Page reports are: Page Views & Unique Page Views. That will give you the optimal view of the data (how many times was a page viewed, how many times was that page viewed by an individual during that visit).

      -Avinash.

      • 67

        Hi Avinash,

        Thank you for your explanation and the hint concerning the Page Views/ Unique Page Views!

        An hour ago I found the specific solution of my page (my quirk): The reason that some visits weren't count as a visit is a "javascript/css expand menu". The GA Code is at the end of the page and everytime the user leave the page to another destination page GA dont count a visit. All Pages with such "javascript/css expand menu" have that problem and all pages without such "javascript/css expand menu" haven't that problem. Maybe I have to put the GA Code on top of the page…

        Cheers

        Peter

  64. 68
    Patrick says

    Avinash,

    I've read through this twice now and it all seems to make perfect sense. I even went through several months of data from Google Analytics for some of our sites and the data confirms what you wrote.

    My question is, why hasn't Google explained this better? I've looked at their definitions of a unique visitor and they come nowhere close to your explanation.

    Thanks,
    Patrick

  65. 69

    Hi Avinash,

    Is it correct to say that the number of unique pageviews for a specific content (one entry in all pages report) is equal to the number of visits to the page?

    Thanks in advance!

    Diogo

    • 70

      Diogo: On paper that is exactly true. The number of Unique Pageviews should match the number of Visits.

      In reality the number will usually be "close enough." Mostly because they are both computed slightly differently.

      -Avinash.

  66. 71

    Hi Kaushik,

    Thanks for such a good post.Can you please explain how the calculations of unique visitors is done if you dont have any cookies or unique id for a user.

    • 72

      Sachin: If the browser does not accept cookies then…

      1. Some tools will ignore those hits completely (so no behavior is recorded)

      2. Some tools will do their best to stitch them together using other signals like IP Address, gap between each hit (to guess that they are for the same session) and other such elements

      3. Some tools will count every hit as a new visit and visitor

      #1 and 2 are more common, #3 is rare. Check with your analytics provider what they do.

      Avinash.

  67. 73

    Now a days Unique Visitors Metrics is renamed with another Name. … Good to see lots of people knows the difference between Metrics and dimensions…. Cheers

  68. 74

    Hi Avinash – really insightful post, as always! I've also just started reading your book (Web Analytics 2.0, which is so incisive!)… thank you.

    I have a question – I am looking at the number of unique pageviews for a particular page. When "page" is the primary dimension with no secondary dimension selected, I get 3,756 unique pageviews. Now as soon as I add "default channel grouping" as the secondary dimension, the number of unique pageviews changes to 3,694 (in this instance, I also get a notification saying "this report is based on 248,323 sessions (87.09% of sessions)").. is reason for a difference in unique pageviews for the same page during the same time period because the former value is not based on sampling whereas the latter is?

    This is quite vexing… would really appreciate your, or another one of your loyal readers input and clarity here!

    Thank you in advance!

    Varun

    • 75

      Varun: When you apply the Default Channel Grouping, or any dimension, the same page might be counted multiple times, this could cause the total to differ (positively).

      In your case, the data also seems to be sampled, which can cause a positive or a negative outcome.

      The role of sampling is to give you an answer quickly. You can learn more about it in this post:

      ~ Web Analytics Data Sampling 411

      Avinash.

      • 76

        Thanks for the response, Avinash.

        I assume that even when segmenting, your response applies.. i.e., the same page could be counted multiple times? To this end, what causes the same page to be counted multiple times?

        When you say, "In your case, the data also seems to be sampled, which can cause a positive or a negative outcome.", do you mean a positive or negative difference in the # of pageviews?

        Thank you once again!

        Best,

        Varun

        PS – thanks for the link to your sampling post :)

  69. 77
    bhola prasad says

    Hi Avinash , i learned the concept in this post but i get stuck with one thing.

    As you say that Google Analtyics does not compute Absolute Unique Visitors when you segment the data, we can get Absolute Unique Visitors for segments by using the "create a filtered profile that just data for the segment" method. I didn't get "create a filtered profile……." part. Can you help me with this ?

  70. 78

    I can see a similar difference in custom events.

    When I view the total events for a custom event by month (Jan 2015, Feb 2015, Mar 2015), the total events do not add up to the same number as when I pull the custom events report for the date range Jan 1, 2015 to Mar 31, 2015. For some of the events, the quarterly number is bigger and for some events the months added together are bigger.

    Any idea why this might be happening? Does this have something to do with sampling? I notice monthly reports are based on a 55% sample and quarterly date range is based on a sample of 16%.

    • 79

      Rohit: There could be a number of interesting things that could be afoot here. I'm afraid without logging into your account and looking at the actual data.

      The best strategy would be to contract with a GACP, they are quite affordable, who can give you very specific advice to help. Here's a list: http://www.bit.ly/gaac

      Avinash.

  71. 80

    Hello Avinash.

    I consider you the web analytics NINJA :)

    I enjoy reading your articles. Analytical, yet simple to understand. Wanted to add that in the new GA "Unique Visitors" is called "USERS".

  72. 81
    William Hills says

    Hi Avinash, I am a big fan of your write-ups.

    Your way of writing is amazing. Really the great information and it explains each and everything in a clear way to a user.

Trackbacks

  1. […] The latest at [Occam's Razor]. […]

  2. […] Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors. – Occam's Razor […]

  3. […] Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors (Occam's Razor by Avinash Kaushik) […]

  4. […] Analytics guru Avinash Kaushik explains what they are, and what the different types you hear about mean, in Standard Metrics Revisited: Daily, Weekly, Monthly Unique Visitors from his Occam’s Razor blog. It’s a long article, but straightforwardly-explained throughout, so do bear with it, and you’ll be rewarded with an understanding of a crucial concept. Most importantly, it will explain why “Daily”, “Weekly” and “Monthly” Unique Visitors are a misleading statistic in most everyday use, which you should avoid. […]

  5. […] Razor – über die Unsinnigkeit von Daily Unique Visitors Wow. Schöne Erklärung, warum man bei seinem Analytics-Package nicht blind jeder Zahl […]

  6. Twitter and the Metrics of the Attention Economy…

    "Ok," your boss says, "I get that traditional marketing is increasingly ineffective and expensive. I also get that we have to try to get on this 'social media' thing, but…" and here's the question that strikes fear in every consultant's heart and in every would-be change agent's heart as well…

    'how do we measure it?"

  7. […]
    Another “problems with analytics” post, this time from Kaushik: Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors. He gives a great run down on cookies, unique visitors, and how to the daily report is useless when looking at periods outside of a day.
    […]

  8. […] Avinash Kaushik gives us the details on visitors and analytics. Read on, as his insights are of great help when digging around your various analytics tools.   Get future posts sent to you free and automatically via email or RSS Feed ! Related Posts: […]

  9. […] Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors. […]

  10. […] 3、Absolute Unique Visitor Absolute Unique Visitor,缩写是UV,是不同的访问者。严格的来讲,是有多少个浏览器客户端,访问了网站。在一个浏览器第一次打开这个网站的时候,页面中的跟踪代码会在持久Cookie中记录,并向信息收集服务器报告有这么一个访问者。只要这个Cookie没有被其他软件删除,那么,在一个报告周期中,这个浏览器不论访问了多少次,浏览了多少页面,访问者始终是一个。Unique Visitor反映的是访问网站的客户端的数目(想象一下一个人既在家用家用电脑也在公司用办公电脑访问网站的情形)。关于这个量度,Avinash Kaushik的文章“Daily, Weekly, Monthly Unique Visitors.”中有详细的论述。 […]

  11. […]
    Segundo problema, hay muchas personas que acostumbran a borrar sus cookies. Dependiendo del periodo a medir, y siguiendo el ejemplo anterior, puede que una misma persona equivalga a seis visitantes únicos distintos (una persona desde 3 conexiones distintas que borra sus cookies cada semana, si calculamos los visitantes únicos de la primera quincena de mes, tendremos 6 visitantes únicos).

    Pero veamos cómo funciona el cálculo realmente en la herramienta. Avinash Kaushik lo explica de manera fabulosamente clara con un ejemplo en uno de sus posts.
    […]

  12. […] 3、Absolute Unique Visitor Absolute Unique Visitor,缩写是UV,是不同的访问者。严格的来讲,是有多少个浏览器客户端,访问了网站。在一个浏览器第一次打开这个网站的时候,页面中的跟踪代码会在持久Cookie中记录,并向信息收集服务器报告有这么一个访问者。只要这个Cookie没有被其他软件删除,那么,在一个报告周期中,这个浏览器不论访问了多少次,浏览了多少页面,访问者始终是一个。Unique Visitor反映的是访问网站的客户端的数目(想象一下一个人既在家用家用电脑也在公司用办公电脑访问网站的情形)。关于这个量度,Avinash Kaushik的文章“Daily, Weekly, Monthly Unique Visitors.”中有详细的论述。 […]

  13. […]
    Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors – Avinash clarifies the differences in grain of various visitor measurements.
    […]

  14. […]
    Top line figures
    According to Google Analytics we had 88,664 unique visitors and 285,522 page views in 2009. Looking at this monthly, as you can see we’re hovering around 10,000 monthly unique visitors:
    […]

  15. […] – Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors. […]

  16. […] Avinash discusses Daily, Weekly and Monthly unique visitors […]

  17. […] Standard Metrics Revisited: #6: Daily, Weekly, Monthly Unique Visitors. – Occam’s Razor – Apr ‘09 […]

  18. […] 3、Absolute Unique Visitor Absolute Unique Visitor,缩写是UV,是不同的访问者。严格的来讲,是有多少个浏览器客户端,访问了网站。在一个浏览器第一次打开这个网站的时候,页面中的跟踪代码会在持久Cookie中记录,并向信息收集服务器报告有这么一个访问者。只要这个Cookie没有被其他软件删除,那么,在一个报告周期中,这个浏览器不论访问了多少次,浏览了多少页面,访问者始终是一个。Unique Visitor反映的是访问网站的客户端的数目(想象一下一个人既在家用家用电脑也在公司用办公电脑访问网站的情形)。关于这个量度,Avinash Kaushik的文章“Daily, Weekly, Monthly Unique Visitors.”中有详细的论述。 […]

  19. […] a quick search though). Avinash Kaushik has a good breakdown of the unique visitor metrics here: https://www.kaushik.net/avinash/2…Also, comparing Analytics Vendors is tricky business as the numbers will never match. Take a look at […]

  20. […] Reading: Standard Metrics Revisited #6: Daily, Weekly, Monthly Unique Visitors /* […]

  21. […]
    Unique Visitors – I just like this one.  I am curious how many new people we are getting into the site.    If you want to read more, this guy had the topic nailed.  https://www.kaushik.net/avinash/standard-metrics-revisited-6-daily-weekly-monthly-unique-visitors/
    […]

  22. […]
    En la gráfica, vemos los totales de usuarios para cada periodo de 1, 7, 14 y 30 días; actualizando la fecha fin del recuento a cada día en curso. Y es que el recuento de usuarios, siempre depende del periodo de análisis.  Por ejemplo, si analizamos un periodo de 7 días, el quien el día 1 es contado como "usuario" (único), al día siguiente – día 2 -, si vuelve, ya no cuenta como usuario adicional, sino que solo suma 1 sesión. Lo anterior es lo más difícil de comprender, el recuento. Una vez entendido eso, lo demás encaja. Para verlo con ejemplos, os aconsejo este post de Avinash.
    […]

Add your Perspective

*