Since the late 1990’s, Unique Vis­i­tors has often been viewed as one of the most strate­gic web met­rics. Count­less com­pa­nies and site oper­a­tors have insisted on know­ing how many unique vis­i­tors came to their site on any given day. Most still do today. In gen­eral, they are moti­vated by the belief that Unique Vis­i­tors rep­re­sents how many “peo­ple” they are reach­ing. Taken a step fur­ther, they believe this allows them to mea­sure how many unique “prospects” or “cus­tomers” they may have.

By con­trast, Vis­its has largely been the neglected stepchild of web met­rics. Most folks know it’s there, but many pre­fer to ignore it in favor of the more pop­u­lar Unique Vis­i­tor met­ric. Gen­er­ally speak­ing, a visit starts when some­one reaches your web­site, and is con­sid­ered com­plete after 30 min­utes of inac­tiv­ity. It is also com­monly referred to as a “session”.

Now for sake of argu­ment, let’s say you can only report one of these met­rics to your exec­u­tives — which is it going to be? Unique Vis­i­tors? or Vis­its? When work­ing with clients, I’m often asked this ques­tion, par­tic­u­larly in the Media and Retail ver­ti­cals. And my answer?

Vis­its…always. Sure, call me crazy — but my logic is actu­ally quite simple.


Here are my top rea­sons for using visits:

1) Vis­its are more accu­rate than Unique Vis­i­tors.
2) Every Visit rep­re­sents an oppor­tu­nity to per­suade or con­vert a vis­i­tor to a cus­tomer.
3) Mea­sur­ing vis­its is based on fairly estab­lished indus­try standards

Now, here are the top rea­sons I would not use Unique Visitors:

1) Unique Vis­i­tors are less accu­rate than Vis­its — Most ana­lyt­ics pro­grams, in the absence of cookie set­ting, fall back on IP address and user agent. This intro­duces sig­nif­i­cant vari­abil­ity in your Unique Vis­i­tor counts and can skew your true site per­for­mance and reach.

2) Unique Vis­i­tors mask your true con­ver­sion oppor­tu­ni­ties — Unique Vis­i­tors are a super­set of Vis­its and may rep­re­sent mul­ti­ple oppor­tu­ni­ties to con­vert a cus­tomer. As such, using Unique Vis­i­tors as the denom­i­na­tor in most per­for­mance cal­cu­la­tions is actu­ally over­stat­ing the effec­tive­ness of your site. For exam­ple, if I visit a retail site 4 times in one week, and pur­chase twice — what is my con­ver­sion rate? If you use weekly unique vis­i­tors, my con­ver­sion rate is 200%. If you use vis­its, my con­ver­sion rate is 50%. Which is a bet­ter rep­re­sen­ta­tion of site effec­tive­ness? Clearly the 50% is much more valu­able in under­stand where your site may or may not be per­form­ing opti­mally. With the 50% con­ver­sion met­ric, I have the oppor­tu­nity to ana­lyze which vis­its did not convert…what hap­pened? Is it a nav­i­ga­tional issue? A cross-sell prob­lem? Or per­haps a remar­ket­ing oppor­tu­nity? If you used Unique Vis­i­tors, you’d never get this visibility.

3) Unique Vis­i­tors are sub­jec­tive — At the end of the day, what is a unique vis­i­tor? Is it some­one who comes daily, weekly, or monthly? How do you decide on a time frame for unique­ness, and why is this time frame bet­ter than any other? Tying into my ear­lier point about mask­ing your true con­ver­sion oppor­tu­ni­ties, the longer your unique vis­i­tor time­frame, the more you effec­tively over­state success.

4) Unique Vis­i­tors aren’t really vis­i­tors — Let’s be hon­est for a moment. Unique Vis­i­tors can rep­re­sent sev­eral things. Opti­mally, a Unique Vis­i­tors rep­re­sents an indi­vid­ual that comes to your web­site. If you have a strong reg­is­tra­tion process, like say an Apple iTunes, you could actu­ally achieve this high level of accu­racy. How­ever, few sites — even in the media space — have such a lux­ury. Rather, most sites rely on cook­ies to mea­sure unique vis­i­tors. Well, we’ve all heard about the issues with cook­ies replace­ments and dele­tions, and some cook­ies are cer­tainly more resilient than oth­ers (did you know there are at least 5 dif­fer­ent cookie types you can use to track visitors?)

Let’s assume 5% of your vis­i­tors delete cookies…that would imply a 5% level of inac­cu­racy around uniques right? Wrong. The fall­back method for unique vis­i­tor deter­mi­na­tion is most com­monly IP and user agent string — a *much* less reli­able approach than cook­ies. This was actu­ally a key rea­son log file solu­tions fell out of favor — because most relied on IP and user agent and hence were highly inac­cu­rate. Because of the inac­cu­racy of user and IP agent, your 5% of cookie reject­ing vis­i­tors can actu­ally skew your traf­fic num­bers by many times over. So you may find that your 5% is actu­ally 15% of your unique vis­i­tors. And because it’s nearly impos­si­ble to rec­on­cile this num­ber (out­side of tri­an­gu­lat­ing with reg­is­tered user counts), you have lit­tle hope in rely­ing on unique vis­i­tors as a true mea­sure of “visitors”.

Fur­ther­more, assum­ing you can set a per­sis­tent cookie, you’re only mea­sur­ing a com­puter — not a per­son. Mul­ti­ple peo­ple use sin­gle com­put­ers. Sin­gle peo­ple use mul­ti­ple com­put­ers. So what is your true unique vis­i­tor count?

To summarize…embrace Vis­its, steer clear of Unique Vis­i­tors
Amidst all this uncer­tainty around unique vis­i­tors, the fact is that I rarely use this met­ric in my web analy­sis and opti­miza­tion efforts. In fact, beware the ven­dor that says they can pro­vide a “more accu­rate” unique vis­i­tor count than your cur­rent ven­dor — until they can solve for the above issues, it’s use­less. Now, there are sit­u­a­tions where I do use unique vis­i­tor counts — and I’ll talk about those in a future blog entry. In the mean­time, my rec­om­men­da­tion is always to use Vis­its — no mat­ter how neglected it may be in your orga­ni­za­tion! And as always, if you’d like assis­tance under­stand­ing how to lever­age web ana­lyt­ics to drive ROI, please do not hes­i­tate to con­tact the Omni­ture Best Prac­tices Group.

9 comments
Marta Silva
Marta Silva

Dear Matt Thanks for all the explanations, as a starter in Digital Marketing it was really helpful.


But I think I'm lacking some functional knowledge. Could you be more specific on your example number 2 concerning the conversion rates? Why is the conversion rate 200% if we use the Daily Unique Visitors? What is the Daily Unique Visitor definition?


Thank you :)

Andy Batten
Andy Batten

I'm late to the party here, but just want to reiterate what everybody else is saying. Great post! I found your examples to be quite effective in getting your point across. Thanks for the post!

dan
dan

I entirely agree. Visits have always intuitively struck me as the best metric for traffic.

Gary
Gary

Hey Matt, Thanks for the interesting read. I used to count uniques, but of course i also realised i need to ultimately aim for higher 'visits'.

Jessica
Jessica

Hi Matt, Being an amateur in analyzing an analytics report, I would like to thank you for sharing such an informative post with us. Cheers.

James
James

If unique visitors are a subset of visits, how would it be possible to have more unique visitors than visits for a given month?

mbelkin
mbelkin

I feel your pain. It’s been a long, emotional-charged debate around the credibility of uniques and I’m glad to hear that you’re fighting the good fight. While there are no silver bullets in winning this debate, I can say that I’m most successful when I challenge people to take action on unique visitors vs. visits. Once they start analyzing the data in the context of visits, they quickly realize how much more actionable and dependable this metric actually is.

Frederic Gilbert
Frederic Gilbert

Hi Matt, Sorry for being late at commenting about this subject but... I totally agree with your arguments about the value of a visit metric over that of a unique visitor, specially today with the avent of Web 2.0. I have used visits as a valuable metric for the last 2 years in order to asses our online results. Because of that, I am always confronted with people around me that are bended by the old rules inherited of mass media marketing as impressions, hits and pageviews,... Thanks for providing me with the credibility of your analysis. Question: are they any chages in your views lately? Bye.