Online advertising visit numbers versus visitor numbers

I wonder if anyone can help me with an industry standard for me to assess a campaign against?

I'm currently looking at a campaign that delivered over 20,000 clicks so given a 20% drop off I would expect 16,000 visits to the site. But we have looked at the number of visitors who came to the site via the unique URLs that were set up and found less than 2,000 visitors (this stat isn't in doubt - it's been double checked and we are an ABCe associate).

We're just checking into visit numbers now and maybe they will be nearer the 16,000 we were expecting but, if they are, this means that the online ads were clicked on several times by the same users. Does anyone have any figures for the average number of times a user clicks on an online ad? I would have expected the average number of visits per user to be only a fraction more than 1?

Thanks,

Ed

Ed Lamb
eMarketing Consultant

020 7598 9976
07879 653608
www.syzygy.net

Thanks Ashley,

What you say makes complete sense - we're doing yet further checking on the numbers now. The number of clicks were provided to us by the media buying agency and I think we can be certain they are correct so we're looking at the areas you talked about now.

We are measuring a visitor by cookie and the time out is half an hour. The timings were again supplied by the media buyers - I am confirming again that these were correct now. We're also checking (again) if we could have missed anything!

One thing I should have mentioned before - there were some PPC keyword buys thrown in within this work. Again I'd still expect a particular user to rarely click on the same PPC placement more than once, but just thought I'd mention it in case anyone knows different. Ultimately, we'll be measuring the PPC work and online ads separately but at the moment I've just mentioned the overall numbers.

Thanks for your help - I'll feed back what the end results are once we know....

Ed

On 12:02:01 4 December 2003 Ashley wrote:
>Hi Ed
>
>If we assume for the moment that you really did get 16,000
>unique users arriving at the site having responded to the
>ad then 2,000 visitors as per your web site stats
>certainly does seem far too low. I don't know what the
>figures are for average_no_times_people_click_on_ad but
>I'd imagine, as you say, it is not going to be much higher
>than 1. There wasn't anything particular about the ad that
>would have encouraged multiple clickthroughs was there?
>
>Of course the response figures may be wrong (did they come
>from an ad server?), but assuming they are correct, then
>may be there are inaccuracies in the site stats after all?
>A few things to think about:
>
>- how exactly are you measuring a 'visitor' (a unique user
>based on IP address, cookie, log in?)
>- what's the session 'time out' set to on the web server?
>- over what time period are you measuring?
>- are you missing / filtering out re-direct pages/sections
>of the site/sub-domains used for the campaign?
>
>[on the above, you may have seen the recent RedEye report
>- http://www.redeye.com/case_data.html?nolnktyp=newsbox -
>which gives a feeling for the sorts of inaccuracies that
>can creep in depending on method/timescales etc.]
>
>Irrespective of the accuracy of the ad response stats, if
>you are measuring near 16,000 visits from 2,000 visitors,
>and this is genuinely only those from a particular source
>(i.e. not mixed in with other site traffic) and if its
>only over a day or so, then surely something is a little
>strange with your site stats? An average of 8 visits per
>visitor over such a short period doesn't seem likely?

It may also depend on the demographics of the intended audience.

AOL use proxy servers, so the issue might be clouded by that.
Many more people are now using built in popup filters in their browser and disallowing cookies.

I'd suggest the media buying data is correct, the measurement of it is wrong, one wrong character in target URL's can throw your data all over the place.

We do a lot of PPC management and track every visitor. Excluding checkbot IP's we normally expect +/- 10% as the sanity barrier.

On 13:17:24 4 December 2003 Ashley wrote:
>And we wonder why conversion metrics (registrations,
>subscriptions, purchases, enquiries etc.) are becoming so
>popular... at least you can be sure of them ;)
>
>Sounds like you're doing everything right so I'd be
>interested to hear if you do get to the bottom of it and
>learn something you could share.
>
>Ashley