Traffic, measurement and the end of page views

Fred Wilson wrote a great post about the increasing problems with page views as a metric for websites.  His words have stimulated two post ideas for me.  Next I will write about widgets and the new challenges for VCs investing in web2.0 companies, but today I want to stay on the topic of measurement.

Much has been written about the problems with page views as a measure of the success of a website so I won’t repeat it in detail here, but in summary three developments all pushing in the same direction are making page views a very unreliable metric for measuring and comparing websites:

  • AJAX and Flash interfaces are delivering more content and functionality on some pages with less page views
  • Atomisation of content – e.g. feeds, widgets and customised home pages – is allowing people to read content without visiting the original website
  • Page views on social networks per hour are very high as people flit from profile to profile

So where does this leave us?

I get asked all the time how I analyse growth in sites and services, and unfortunately there is no easy answer. 

Alexa gives a useful ready reckoner abut where a site is and what sort of trajectory it is on, but is a bit volatile and unreliable as a measure.  It is free and very convenient though, so that is the first place I look.  If a site has a high ranking (popular sites have low rankings) then it is unlikely to have much usage.  Google rank is another (very) crude measure that is free and convenient.

Moving beyond that there is a basket of measures that are available and different combinations will be appropriate for different sites.  Page views is certainly one of these measures (no matter how much AJAX or how advanced your widget strategy if your service is popular you should be getting some hits..).  Others include time on the site, number of registered users, number of subscribers, number of times people come back, page views per member and contributions per member.

Figuring out which metrics are important is critical – the key thing is to demonstrate your site is getting used a lot, by the right type of users, and that usage is growing.  If your business model is advertising then volume related metrics are going to be up there, but if you use a lot of flash then you will be looking more at time on the site than page views.  If you are about subscriptions then acquisition costs, subscriber numbers, conversion rates and churn are likely to be your key measures.

My advice would be to figure this stuff out and put it on your site.  Admob do a great job of this – their key metric is ad views (they are a mobile ad network) and they put the number right up there on their front page.  Putting your key metrics in the public domain shows a confidence that is appealing to investors and customers alike.  It also puts you in charge – you decide the metric rather than having to explain why page views isn’t right for you.  Going through this process with sufficient rigour to publish will be a good discipline as well.

If you are a young company (and you have a good story to tell) then this can be your best way of getting noticed.  You can’t do this stuff when you too small though, and judging the point when your numbers and growth go from sounding tiny to sounding exciting in the context of a new market is part of the art of marketing.

  • If you are a young company (and you have a good story to tell) then this can be your best way of getting noticed. You can’t do this stuff when you too small though, and judging the point when your numbers and growth go from sounding tiny to sounding exciting in the context of a new market is part of the art of marketing.