What and how to measure Social Networking websites

By social networking websites, I am talking about the likes of Facebook, Myspace, Linkedin, Xing and StudiVZ (and of course any other web property who are engaged in Social Media activities – like any decent blog you know).

Social Networking is the prime headline for Web 2.0 and highly successful so far. This fact only leading towards an even greater responsibility in trying to measure ROI on ones social networking activities, especially as it is somehow agreed upon that only a limited number of players can win (having Metcalfe’s law in mind). Remember how quickly Friendster somehow lost momentum, I bet you that they did not deploy detailed analysis, at the level required in this game, within their organisation! – if so, they would have spotted the decreased user engagement and its relating metrics in time.

First I would like to conclude that one have to accept that however sexy social networking is as a trade, there MUST be a set of measurable KPI’s that are strongly aligned to ones business objectives, like any other organisation that operates with the web being a channel of theirs. Hereby concluding that KPI’s are meant to be acted on and not just reported on. Most of the social networking sites that I have been working with somehow work to optimize the following two business objectives.

Social Networking business objectives:

  • Increase Advertising and/or premium member-ship Revenue
  • Increase User Engagement

Now that we have clear business objectives defined – and some that in my humble opinion are pretty much spot on in regards to the overall success of any social networking business. Assuming that you agree with me, we now have the opportunity to define a set of specific KPI’s and accompanying important metrics.

Social Networking advertising revenue KPI’s:

  • Advertising Revenue
  • Visits per week
  • Ad units per visit*
  • Ads served*
  • Ad CTR

* The “Ad units per visit” KPI is a replacement for the oldpage views per visit” as a way of indicating the size of your Ad inventory and the “Ads served” KPI is the actual number of revenue generating banners or other media type served in those units.

Remember, that when talking about these 5 important KPI’s (and no business should really have more than 5 KPI’s on a management level) – we are not just talking about a set of basic reports on their performance in retrospective – we are talking about collecting data on a granular level so that we can both report AND more importantly do analysis on these – by segmenting, slicing and dicing them any way imaginable.

In the end for any Social Network to succeed and survive it’s users have to be engaged with the website – and that engagement actually turns into positive revenue as they spend more time, take more action and convert on a higher rate than non-engaged users. Concluding that we are looking at a win-win-win situation for the user, the social network and finally it’s advertiser. We of course have to agree on what “user engagement” is to begin with and inspired by Eric Peterson I would say:

Social Network User Engagement is an estimate of the degree and depth of visitor interaction on the website against a clearly defined set of measurable goals.” – and with that in mind we can define our KPI’s.

Social Networking user engagement KPI’s:

  • User Engagement*
  • Anonymous visitors to members conversion rate*
  • Active member length
  • Time since last login
  • Total time spent on site

* The “User Engagement” KPI is a custom session metric designed and calculated from a set of basic metrics such as for example; pages viewed, time spent on site, time since last login, comments or other content submitted, subscribed to a feed or alert and so on – the calculation of this KPI is highly dependent on the structure of the Social Network in question). The “Anonymous visitors to members conversion rate” is very much a growth KPI that is used until the Social Network reaches its final plateau.

I think it is of the utmost importance (as in any other media company) to understand what benefits the advertiser and how he get the best possible ROI. Social Networks have the opportunity to create value for the advertiser far beyond the click — if the Ad creative and Ad content showed allow this — so it is also important to track the actions taken beyond the click (and in general for the website). Such accompanying social network user engagement action metrics like:

- Forwarded advertiser content
- Endorsed advertiser content
- Rated advertiser content
- ..and other obvious types of engagement with an advertisers content

As you can see the secondary audience is as important as the immediate audience and somehow one should be able to illustrate that value to ones advertisers – as in being able to increase revenue based on the fact that the advertiser gets proven value increase by word of mouth with the secondary audience derived from the immediate audience.

I also find it important that one set up specific custom fields for tracking such accompanying social network user engagement metrics like:

- User profile completion level
- Premium member level

Both the accompanying set of metrics for advertiser and user engagement are super specific – so the above two small lists are only a couple appetizers. More importantly how do we “work” the KPI’s

Social Networking KPI control elements:

  • KPI Targets
  • KPI Indexes
  • KPI Competitive Intelligence

I am sorry to add to the list of tasks, but one simply have to create a credible target for every single social networking KPI that we discussed above (credible as in utilizing competitive intelligence to set realistic goals) so that there is a clear driver within the organization on where it is going. At the same time while working towards those targets one need to create a sensible social networking KPI index that can be used when doing analysis for the KPI in question (sensible as in taking into consideration how the KPI fluctuates due to season, campaigns and other factors). Finally utilizing competitive intelligence to spot market opportunities for optimization and general KPI improvement.

NOW – having clear business objectives, 10 well defined social network KPI’s and a set of KPI control elements in place – we get the opportunity to put all this to work, as in nobody should go to the extent of setting up a framework as the above without aggressively pursuing a performance improvement in ones business objectives. It’s really not worth measuring something if you cannot or will-not take any action on it, and to be worth taking action on, it has to have some kind of measurable monetary value. There is no magic 7 points to optimize, but given the framework as described above one would have an unlimited number of opportunities to make more money! – find a couple of suggestion below:

Social Networking optimization opportunities:
(7 basic and to some extent obvious suggestions)

My first comment is that before trying to optimize anything for the better – some sort of potential monetary valuation should be put on the effort as in; are we looking to increase revenue EUR 10.000 per year or are we looking to increase revenue EUR 2.000.000 per year and with what certainty can this be determined.

As an example, let’s say that we have 15 Ad units served per visit and that we have an Ad unit CPM value of EUR 3 and that we have 1.000.000 visits per week (notice how all of these are KPI’s) – this leaves us with a weekly Advertising Revenue on = EUR 45.000 (1.000.000 visit * 15 Ad units per visit / 1000 (CPM) * 3 EUR). The optimization opportunity in question estimates that we can increase the number of Ad units served to 18 by adding a small additional unit AND that there will be a 0.30 EUR decrease in Ad unit CPM value due to this (notice that we utilize out KPI’s index here). Therefore an monetary opportunity on EUR 3600 per week or in perspective approximately an opportunity per year of an additional EUR 188.000 in revenue!

Here goes is my 7 basic and obvious suggestions (very specific and less strategic though) on what you should look into when optimizing a social network:

  1. Segment users based on their community activity (this could be anything from viewing pictures or videos to commenting and participating in forums, this is essentially all the defined actions in your web analytics package) – and list this according to the Ad CTR KPI and comparing the different segments to your Ad-CTR index. This giving you an opportunity to do better internal promotions (as basic as a link in the right place) for better converting content (that above your index) on poorer performing content pages.
    Opportunity: To increase the Ad CTR and thus the Ad unit CPM value.
  2. Segmenting the Anonymous visitors to members conversion rate KPI by campaign channels (in my world ALL incoming traffic should be part of a campaign, this including SEO activities) to find and focus on better performing channels and campaigns.
    Opportunity: To increase the Anonymous visitors to members conversion and thus if visits are constant the number of new members
  3. Perform single-page FORM Analysis (tracking every single FORM field) on the sign-up page and cross-reference this to the Anonymous visitors to members conversion rate KPI – getting to know which fields make Anonymous visitors abandon the sign-up form.
    Opportunity: To increase the Anonymous visitors to members conversion and thus if visits are constant the number of new members
  4. Create a custom report with the metric Time since last login KPI and a grouping by entry pages. Filter out all the users who have visited the site the last 5 weeks, so that you end up with a list of somehow disengaged users, who returned to the site. The most successful entry page for disengaged users should be the one heavily used in e.g. Email Marketing. (A common page on this list is the “Your Friend x added you as a friend”).
    Opportunity: To decrease the Time since last login and thus increase the numbers of Visits per week and thereby increase Ads served and finally increase Advertising Revenue as a result. At the same time increasing Active member length.
  5. Run an out of the box standard Action Participation report and add the User Engagement KPI – creating an easy an instant understanding of what content triggers a passive visitor to become a active contributing user and member of the community. As of now you probably do not really know whether the highly engaged users connect to the community by watching user submitted videos, read gossip, chat with friends or what have you.
    Opportunity: To increase the User Engagement compared to our index and thus we typical see an increase in Visits per week and thereby an increase in Ads served and finally increased Advertising Revenue as a result.
  6. The Total time spent on site KPI is not only important because Nielsen//NetRatings has officially replaced “page views” with the “time spent” metric for their official ranking — whether you agree with Nielsen or not — it is still very important to know which content groups that users spend their time on – the most valuable resource they have. Create a report that shows Total time spent on site divided per content group and conclude the individual standing on these content groups weighed against your competitors using competitive intelligence. Not as a relative result, but as in looking at the fact that if a user is willing to spend 15 minutes watching user generated videos in general, this is the actual number to target). Optimizing content groups that deliver below industry standard time consumption.
    Opportunity: Increase Total time spent on site and thus an increase in Ads served and finally increased Advertising Revenue as a result.
  7. Try to increase the “quality” of your Ad units per visit KPI by using External Data Sources and relate your internal user database to the Member-ID tracking done by your Web Analytics package (this will give a plethora of opportunities), but one obvious is generating a simple report by content area and thereby get insight into the demographics and sociographics for every specific content area (network of people).
    Opportunity: Increase the Ad unit CPM value by demanding higher pricing for better targeting and at the same time expect higher Ad CTR

..and a bonus optimization comment/suggestion inspired Jim Novo on why using the “User Engagement” KPI on a management level can save your business (or Job). Let say you have a Social Network for students and the number of members and total number of actions in general is increasing (up and to the right), but the “User Engagement” KPI and the related “Time since last login” is decreasing in value more rapidly – then at some point the disengaged users overpowers the increase in members and the website starts to spiral downward in volume. One can predict this downward spiral will happen by monitoring these KPI’s (even as a basic management Alert in your Web Analytics tool) so one can try to take action before it’s too late. Most people look at volume as a measure of popularity and growth and by the time volume starts dropping, it is already to some extent too late to save the website. The audience has already disengaged. The user engagement drops off before the traffic volume does, and that is why user engagement is predictive and directly addresses future value.

Even though these 7 basic suggestions + bonus comment (and both you and I could come up with 50 others) somehow seem simple – I am confident that anybody running a social network would be far better off looking at and concentrating on the KPI’s framework described.

My overall conclusion is that one cannot be successful — in the long run — running any social media activities if not one or more activities around enterprise optimization analysis is deployed – and I hope that the above somehow indicated and inspired which direction to go in.. sorry for the partly long post (essay) :-)

  • Clive Fernandes

    That was a fantastic post. I’m currently working on KPI’s for a new website being funded and you’ve given me a lot of information to dwell on. Thx a million.

    Wouldn’t some sort of brand recognition metric be useful in the analysis of a SNS. My view is that the internet needs to move from just CPA advertising to selling branding oppurtunites like the mainstream media.

  • Clive Fernandes

    I’m a bit confused what you mean by “Active member length” and what would be the method to measure that. Could you please elaborate.

  • Dennis R. Mortensen

    Hi Clive,

    Thanks a lot for the positive feedback.

    I do agree that “some sort of brand recognition” is worth looking into. However; I do not see it as a KEY performance indicator, but definitely useable in e.g. pricing Ad-views.

    >>I’m a bit confused what you mean by “Active member length” and what would be the method to measure that

    I think it is extremely important to know for how long (in time) your members stay active. I would in a simple way measure this as: “last sign-in date” minus “sign-up date” (this does not take into consideration potentially disengaged periods. There is a long debate on how to handle this… which is for another post. :-)

    This KPI is valuable in many ways… one obvious one is – how much you can spend on traffic-acquisition (and essentially member acquisition). If you know that you average Active member length is 17 months. Think of this KPI as being equal to an e-commerce site using CLT (Customer Lifetime Value)

    Cheers mate.. :-)

  • Pingback: VisualRevenue | Online social network participation inequality and the impact on the User Engagement KPI

  • Pingback: VisualRevenue | Your most popular page is likely NOT to be your most popular page

  • http://www.techlinks.net Jessica Hart

    Let me begin by saying WOW! I am currently working on a business plan for a new revenue model for an exisitng Technology Community website and I m stuck in the “chicken vs the egg” paralysis by analysis mode. I know just enough to be dangerous (ironic given the nature of our site’s community :) .

    The existing site has very little tracking/reporting in place, no online ads and essentially no way to measure ROI for our clients who currently pay an annual sponsorship fee to have their company profile listed and the ability to upload community publishing and share events/announcements.

    My mission is to retool our site and create a suite of new services that leverage “best Practices” in social networking, online ad placements, content sponsorship and contectual marketing services.

    I feel like I really hit the motherload of information with this post – I am humble by your supreme intelligence Dennis…I totally “get it” and am prepared to execute accordingly – what I need advice on is how to establish some basic valuation of my exisitng site and translate that into new revenue forecasts using either a flat rate or CPM model to determie appropriate pricing for the new offerings.

    I see from these post dates that I am about six months late to these conversations but do hope I can respectfully call upon this group to help me get back on track.

    With Appreciation,


  • http://visualrevenue.com/blog Dennis R. Mortensen

    Hi Jessica

    Thank you for all the nice words. In regards to the specific business model (merchandizing, pricing, etc.) – I think I would be a tad too cocky if I recommended you going left or right, I however believe that given that you collect the right data and do the right analysis – that you will be able to conclude how to maximize revenue!

    I’d personally never been to fond of flat rate as it gives little room for revenue acceleration, but that’s just me :-)


  • http://www.techlinks.net Jessica Hart

    Hey Dennis,

    Thanks for the post. I too would prefer a CPM model to a flat rate – I just don’t know if I have the traffic to support it right now. Based on my current site stats, I would be looking at a CPM rate of $225. Since we are revamping the site to an industry specific professional social networting site with workflow tools built in to accerlate peer to peer and peer to vender connections I know I can shoot for a 3 digit CPM rate (Linkedin is at $75 CPM and national industry community sites in my niche boast $175 CPMs – I feel like I should stick with a flat rate until I boost traffic.

    Thanks so much for responding.


  • Pingback: VisualRevenue | Online Video Analytics - KPIs

  • http://tshawkins.blogspot.com Tim Hawkins

    I think there needs to be some metrics that surface the success of the SNS. some suggestions are:

    1. Connectivity – The number/percentage of people with 2 or more connections to others. Possibly expressed as first, second and third degree connectivity. This could be expressed in bands, ie connectivity above 2 contacts, 5, 10 25, etc…. It would determine who your most socially active segments are.
    2. Density – The number/percentage of connections vs members (strictly speaking the density should be expressed as the a percentage of the total number of possible connections, but in real life SNS that number grows impossibly large very quickly as the member base expands, and the percentage shrinks fast).
    for example if there are 4 people in the network and two have contact with each other, then the density is 50%. Note it is quite normal for density in this fashion to exceed 100%

    3. Interactivity, the percentage of interactions, (messages, writes, links) vs the Density.

    If a SNS system is not exhibiting growth in these areas then it is behaving as a simple flat site, and engagement will be low. Which will have a negative effect on the other KPI’s listed above.

  • Pingback: VisualRevenue | 27 most popular Web Analytics blog posts of 2008

  • Pingback: Popular Web Analytics Post of 2008 « SEO Ahmedabad: A place to learn & share all about SEO

  • http://www.totspot.com Michael Broukhim

    This is a phenomenal post. Thank you for writing it! Curious to know, do you suggest most of this analytics capacity be developed internally? It seems like some of the KPI’s you’re suggesting aren’t easily trackable with any sort of off-the-shelf analytics package (e.g. Google Analytics)… would love to hear any thoughts you have actual data collection methods.

    For instance, how would you go about creating this report:
    “Create a custom report with the metric Time since last login KPI and a grouping by entry pages.”

  • http://visualrevenue.com/blog Dennis R. Mortensen

    Hey Michael,

    I am neither for or against an internal driven analytics team, I am however a big fan of moving towards the TRUE data driven organization. AND I find it difficult to see that happen without having at least an internal expert and an advocate of a data driven decision model on the ground (making sure that processes are built around those exact KPI’s and that actions are inevitably based upon those)

    Some of the KPI’s aren’t directly reportable within e.g. GA as you so correctly note; But not only that, they are not directly collected as data points to begin with. (you can fake it with some decent proxies if you want, but… not really)

    >>For instance, how would you go about creating this report:
    >>”Create a custom report with the metric Time since last login KPI and a grouping by entry pages.”

    I would set up a custom field called “Time since last login” which you populate (the measure being days) on each login, whether that being a manual login or an automatic cookie login (you will get the number of days from your “user login log”, which you of course need to have in place in your internal user database). This will collect the data and you can then start play around with it. Some of that play will unfortunately (and I am saying that with great regret as a vendor… argh) have to be done with exports and a bit of Excel from time to time.

    .. but all that said. You are in NY, right? You are more than welcome to pop by and get a free diet coke at Yahoo! on 18th street – and have a chat about Analytics and Startups.

    Cheers :-)

  • http://www.optimizeyourweb.com Kristjan Mar Hauksson

    Just re-caping on this post, great read

  • http://visualrevenue.com/blog Dennis R. Mortensen

    Thanks Kristjan.

    I wrote in on a beach in Singapore!


  • Pingback: Visual Revenue | What and how to measure Online Finance websites

  • Pingback: Réseaux sociaux : vers de nouveaux indicateurs de performance ? @Economie Numerique – Blogue du cours

  • Pingback: Measuring ROI « Diary of a Gen Y Professional

  • http://ultralightstartups.com Graham Lawlor

    Hey Dennis,
    Another fantastic post. I’m proud to be the first commenter on your timeless treatment of this topic in 2010 – the 4th year running this post has received comments. Obviously the idea of data driving business decisions never goes out of style. Congrats!


  • http://www.youtube.com/watch?v=QO0ORWnPc2k Done for you Traffic

    My brother recommended I would possibly like this web site. He was once entirely right. This publish actually made my day. You can not imagine simply how so much time I had spent for this info! Thanks!

  • Pingback: Social Networking, the fabric of global internet | Web Analytics India Blog