Is the page view dead?

There are two reasons that make it increasingly difficult for page-view to be a good metric to compare across sites. First is in-page ajaxy activity. For example, on SlideShare you can download powerpoint file, add slideshow as a favorite and comment without going to another page. The second reason is related to consumption of multimedia content such as video and slideshows. A user might watch a slideshow for a while without impacting page views in any manner.

What can replace the page view? Neilsen moved from page views to time on site as a core metric last July. Now they say that the impact has been greater for video sites than sites with a lot of ajax. A new web metrics startup Nuconomy measures things like number of comments, downloads etc. to track engagement.

At SlideShare, we are starting to use “No of clicks per visit” as a measure of engagement. I like it more than time per visit, since it gives an index of activity on the site. At the same time it takes the multimedia viewing and ajaxiness of the site into account.What do you think about “No of clicks per visit” as a metric?

11 thoughts on “Is the page view dead?

  1. Page views or unique visitors are bad measurements because they focus on abstract metrics instead of the valuable customer behavior that generates those metrics.

    So, they focus on the check they get at the end of each month, instead of the reason why they get that check. 1 million page views might mean 1 million dollars, but if you’re not converting any users, then next month, you might not make a dime.

    I’ve been trying to push a meme at work around ‘behavior-driven metrics’ with the gist being you have to measure the user behavior that drives the money.

    So, for a site at work, the business model relies on return visitors. Where the business wants to focus on unique visitors per month, I’m trying to push them to focus on return visitors per month.

    It seems like an obvious answer to me, but in the end, the business is on the hook for delivering so much money a month. The end-all, be-all for most of the business side will be a metric that tells them how much that check will be.

    So, my latest thinking is to go for a compromise: track engagement (the user experience) and monetization (the business success).

    * page views per session (engaged) AND ad views per session (monetization)

    * video ends (user engaged) AND ads views per video (video monetization)

    * return visitors per month (engagement) AND ad views per return visitor per visit (monetization)

    This is still a problem I’m working through, so I’m kind of thinking out loud here…

    I think number of clicks is better than time on site, but the best metric measures the user behavior you believe is critical to the success of your business model.

    Is that number of uploads? Do you believe uploads drive the referral traffic that’s key to your ad revenue?

    Is the key behavior number of registrations? Number of slideshow starts? Slideshow completions?

    Time on site, number of clicks, unique visitors, whatever. None of those matter if none of them contribute to what makes your business run.

  2. Austin, Rashmi:
    It seems that you both put a lot of thinking to the subject of measurement behind page views.
    I’ll be happy to show you what we do with some of our customers and get your feedback on it.
    You can reach me at shahar at nuconomy dot com.
    Shahar Nchmad
    NuConomy, CEO

  3. Rashmi,

    Your solution is an interesting reaction to unease with the page view. As a measure of viewer engagement, it sounds very compelling, but as Austin mentions, it only measures one part of your picture.

    Just remember that you get more than one measure for your current success. For slideshare, a combo of uploads (supply), click per visit (engagement), and one other about ad impressions (monetization) gives you a simple snapshot of the overall picture over time.

    In my mind, the key is to keep doing your work, checking the snapshot on occasion to make sure that you’re still going in the “right” direction based on your strategy, and drilling down into details when you see warning signs or hit the motherlode.

    It will be great to hear how it goes!


  4. Rashmi,

    You are essentially trying to measure the user interest in your site, but then you have different kinds of users. I have thought briefly about in another context and here is my half baked thinking… this is all theory but am pretty positive this is in right direction…

    Define user segments, you might already have done personas.. in any case.. one way could be publishers, consumers (might be a separate consumer who is download heavy) , by standards/occasional consumers

    The goal of each of these guys is different but connected. The publisher’s goal would be to offer great content and gain visibility.. so how do you measure if content is great or not – one way is star rating, another could be sharing of links by other users (more thinking can draw more ideas), similarly visibility – if you track who is getting how much visibility (in terms of views) and then map these two, you’ll see how successful are in you in helping publisher achieve her goal.

    Another matrix you could build on top of this is what I call “Zero to Sixty” – as in acceleration measure (like in cars) is amount of time it takes for someone to visibility, may be that 10% presentations on on reach 10,000 viewership in 30 days

    For consumers – they care about quality of content, so you can start taking the above matrix and looking at it at site level and see if the quality of content is improving.

    The other part with consumers is presentation completion rates – how many reached to the end. And second aspect is how much time.. a parallel example to this is market research using online surveys, there are mathematical models available to find outliers (by measuring who run through surveys very quickly etc..). You could explore similar models to see what is the amount of time one should spend on online presentations for it to be valid.

    These could be interesting stats to include in product blog as well….

    Similarly with other user segments.. this really tells you how successful your product is in meeting customer needs which will drive business value..

    The business value in the end will get measured on what parameters affect growth for the business.. is it more people, is it repeating customers, it is amount of new (good quality) content being generated every month etc..

    – Alok Jain

  5. Rashmi,

    I think a standard metric may not apply to different types of website that we have now. Eearlier it was all about consuming the htmls (pages) and hence the page views. But now the user behaviour is different across different web sites – social networks, blogs, video and other media sites. For each of these segements there has to be different qualifier,not just for quantity but quality also. In future we would see these sites reporting it in some other terms and may be trough a set of metrics instead on one or two.

    Still the majority of the web would use Page Views as their main metric, bcos they are still about the pages. The metric has to do more with ‘user behaviour’ and the ‘page behaviour’.

    In case of slide share – click per user would be a top reporting metrics and for internal consumption I am sure you would be tracking much more
    – How many user actually go through slides (time, start to end)
    – The number of downloads (personal benefit)
    – Bookmarking, forwards, etc (social benefit)
    – Comments (content popularity)


  6. Pingback:»Blog Archive » How Web2.0 has all but killed the pageview?

  7. There can not be a single standard metric for comparison of all catgories of websites. For example, clicks can’t be used for a comic website, something like xkcd. Similarly page views cant be used for AJAX based sites as you mentioned. It’s all about competition within a particular category of web content, you have to choose a gauge that suits your needs the best.

  8. No of click per visit -hummmmmmm Okay if i will visit your site for the first time then no. of click you will get from me will be high -as i will be keen to surf it to know what there to offer me then when i will be the unique visitor my activity of the site may be lessor as i will clcik on to the new stuff only -What say?

  9. Hello Rashmi,

    The idea is great… but the chances of people mistaking it for page views are quite high!

    I noticed that not everyone goes through all slides in a slideshow some quit mid-way (I observed this trend among my friends) So the amount of slides viewed in a slideshow can be used as a rating system of sorts.. so the “uploaders” too get to know which part their presentation turned ‘boring’!

  10. Hello Rashmi,

    First, congratulation for all your work and your blog.

    I hope you can bring some fresh air to that Metric holly grail.

    I undertsand that designers, webdevelopers and marketers are pressed to find THE indicator that will help solve all the budget problems.

    It would make sense for slidesahre to measure clicks as for each slide you will click. Maybe not for youtube. What i would like to see is measures that matters to the business objectives and page views, minutes per visit, etc are not necesarely the one that would indicate how healthy (wealthy?) is a business.

    I think that we should be able to understand what is really relevant for the business and see where does it fit in the technology stacks.And then measure it.And not starting by a measure and than figuring out how it fot in the business.

    People come to slideshare to view slides so maybe the number of click on next slide is a key indicator for slideshare.No t all clicks , only the next slide button.

    The analysis of that single data could be a very good measure to see where the site is heading.

    What is the expected growth rate of that indicator. Is it growing at the expected speed. This would be like the heartbeat of slide share.

    If that heartbeat doesn’t perform well then you could have a look toblood pressure, tension, and other health indicators.

    So in websites, Idenfying this heartbeat doesn’t necesarely has to be the same for all even many people try to convince us that it must be like this.


Comments are closed.