How balanced is your online media mix?

A benchmark compilations comparing average media mix across websites

Naturally all digital marketers want more (quality) traffic. This begs the question where do we find this traffic? What is our optimal mix of online media? One approach to answering this question is to turn to the traffic sources report in Google Analytics.

This will show you a pie chart showing the relative importance of your different traffic sources. If you find one source of traffic like direct traffic or search traffic dominates, then there may be an opportunity to work to build other traffic sources like using out reach to other sites to build referral traffic. There’s a risk if “too many of your eggs are in one basket”, particularly as Google changes it’s algorithms…

How does my traffic mix compare?

Even once you’ve reviewed your traffic mix, there’s always the question of how does my mix of traffic sources compare. Well, thanks to the recently published Google Analytics Benchmarking Newsletter you can now get an idea of how your mix compares based on the global average below:

Note that this was published mid 2011 and as of 2013 Google hasn’t published any other more recent benchmarking updates although they say they will recreate the benchmarking service.

If you’re not familiar with these, the four main traffic sources reported at this level in Google Analytics are:

  • Search traffic – This groups both natural and paid search (Adwords)
  • Referral traffic – This is traffic from other sites which have direct links to your site
  • Direct traffic – Direct traffic results from URL type ins, bookmarks or when email marketing isn’t tracked. These days direct traffic will also include non-browser traffic from visitors clicking on apps for reading social messages like Hootsuite or Tweetdeck or other mobile apps linking to a site. See the details from Tim Leighton Boyce in the comments for why “direct traffic” isn’t really an appropriate term today or in his excellent post “If social media marketing is so big – where is all the traffic“.
  • Other/campaigns – Campaigns include Adwords when linked to the Google Account and any other campaigns like affiliates, display ads and email campaigns when these have had marketing campaign tags attached. In this compilation Adwords is included under search engines.

Bear in mind that the mix will vary dramatically according to any number of factors, that’s the problem with averages… In particular large brands tend to have a much higher proportion of direct traffic, often >> 50% in my experience as people enter the URL direct. You’ll also see that direct traffic is higher if email campaigns haven’t been tagged when they are then added to the campaign or other category of sources.

Site traffic mix according to Outbrain

Another view on the mix is provided by Outbrain who provide sharing widgets and content recommendations for publishers. Although may be are embedded on smaller blogs, they’re also used by many large online publishers. So there is a greater reliance on direct traffic, but they have also identified traffic direct from social and content-related sites. It’s noteworthy that direct social traffic is relatively low (of course there is also social traffic from apps which will be included in direct).

Here is a breakdown of referring sources in terms of the type of referrer represented:

Excluding direct and in-site traffic, an adjusted breakdown of traffic share is illustrated by the following pie chart:

How does your mix compare?

  • Tim Leighton-Boyce

    Please can I pick up on one point of detail. The standard description of ‘direct’ traffic, which is repeated here, is far short of the truth. These days a lot of social traffic coming from clicks in mobile or desktop apps will also be attributed to ‘direct’. This is a growing problem.

    I’ve written about it and shown a recent small sample of data here
    with links to Thomas Baekdal’s bigger data set from a year ago.

    Direct is an extremely fuzzy set of visits in Google Analytics these days. To make matters more confusing it is treated differently in Multi-channel Funnel reports compared to traditional GA attribution reports.

    I wish they would re-name it as “unattributed”. That would be a much more descriptive word.

  • Dave Chaffey

    Thanks very much Tim, for the reminder about some social media referrals being including in direct traffic – I have updated my explanation of direct traffic. That could be quite a big detail for many sites who are active in using social media to drive traffic.

    Your article and link to the other data set is interesting. I took a look at our data from Hootsuite shortened URLs and see that around 60% of our visitors from social media aren’t direct from the social networks as referrers, but from apps and other sources. So those would be wrongly recorded in Google Analytics.

    Yes, “unattributed” would be more accurate, but in the meantime your comments have helped explain what “direct” traffic may include.

  • http://[email protected] Paul

    Hi Tim, to avoid or at least minimise this issue, best practice would be to tag your links and then shorten them. Then you’ll get a more accurate picture of your traffic sources and more accurately categorise the “direct” traffic.

    Google Analytics and Site Catalyst have traffic source tagging parameters, eg:

    • Dave Chaffey

      Tim is certainly aware of this, but your comments will help others who read the post – so thanks Paul!

      We have more guidance on campaign tracking in Google Analytics and how to segment different sources here:

      We ran a poll that shows that tracking using parameters for social media is relatively rare (16% of respondents).

      I’m not sure which group the new “social sources” in GA are placed – – I guess the “other” group?

    • Tim Leighton-Boyce

      Hi Paul, I take your point about tagging the links which you post yourself. But that aspect of ‘social marketing’ is a more sophisticated variant of the traditional ‘send out an email blast’ and not what I’m anxious about.

      I suspect that the hidden power of ‘social’ lies in the amplification effect when someone spontaneously decides the mention your excellent service or offers and that is picked up by others. I would like to say ‘real power’ as opposed to ‘hidden power’, but my point is that the data is confusing and may be deceptive.

      The kind of effects I’m talking about are the unexpected spikes which come when someone posts a link to an offer on forum like Monesavingexpert. Those tend to include referrer information and so they can be seen in GA, even though the link is user-generated and contains no tags.

      If the same thing happens on Twitter, or in another context where apps are a common way of viewing the link, we know almost nothing about it.

      If we could measure and value the benefits of that kind of spontaneous vote of approval — and not just relating to offers — I believe it could be used to prioritise improvements in user experience, content and customer service.

      That sounds like a good thing to me. I’d just like to be able to prove it!

  • Pingback: Smart Insights guest blog – Using web analytics to benchmark SEO performance « digitaljuggler()

  • Pingback: Using web analytics to benchmark SEO performance | Digital Juggler()

  • Pingback: Using web analytics to benchmark SEO performance - Digital Juggler()

Feedback Form
Feedback Form