Back in October 2011, Google made the switch for SSL search for users signed into Google accounts which meant the loss of keyword data from organic search. The visitors would be not classified as (not provided) within platforms such as Google Analytics. I have been watching how the NP traffic has been steadily growing across a number of websites but have also been monitoring how those types of visitors interact when they visit the website.
To get a decent sample size of Google Analytics data I looked at a period of time from the 1st November until 28th January with 2.4 million organic visitors. The first big observation is that that while there is a strong preference for tech focused sites to attract more (not provided) traffic it is not always industry specific.
My research still shows a bigger influence on how many (not provided) visitors you will see in your web analytics reports still seems directly related to the total number of US visitors your site attracts. So if you currently have a fairly low percentage of your audience from the USA, (not provided) will currently be less of an issue for your business. I can see the problem would be if your website depends on organic traffic from the USA you will continue to see the issue of (not provided) grow during 2012 based on early adoption statistics of Google+.
With my sample size it averages out to be around 5.02% of total visitors during that period are classed as (not provided) which is slightly lower than I expected.
Can you monetise (not provided)?
The big issue I see based on my query data related to (not provided) is that the Google AdSense platform doesn’t seem to be able to monetise this traffic that well. The initial issue is that even though (not provided) brings in 5.02% of site visitors, it only accounts for 3.16% of AdSense revenue. The other issue is that the average revenue per organic visitor is $0.0051 but only $0.0031 for (not provided), so you are making 40% less per organic visitor if your site relies on platforms such as AdSense for your primary revenue stream.
What does this mean for you? You are losing around $0.40 out of every possible dollar of revenue if the visitor is from the (not provided) audience.
Does (not provided) convert?
If you are running a lead generation website or an e-commerce store the (not provided) traffic looks to be a far bigger issue as the data set shows (not provided) performs even worse for conversions than other organic traffic. The data shows an overall organic traffic conversion rate of 0.97% compared to (not provided) of 0.67%. My analysis showed the other issue is that even though (not provided) accounts for 5.02% of traffic it only accounts for 4% of organic conversions.
What does this mean for you? You are losing around 31% of your potential leads/conversions if the visitor is from the (not provided) audience.
What can you do?
You should start to examine what are other options for monetising (not provided) traffic such as trialling different ad networks, offers for email signups and even in-house advertising. It’s important that you start to look at your web analytics data to see of the (not provided) visitors never convert or engage with your onsite advertising.
Also see;
- The hidden value of Google (not provided) for SEO
- Making sense of Google (not provided) data
- Dear Google: This is war
Hi David
Possibly a few wrong conclusions.
Not Provided are early adopters
Were early adopters equal for conversion previously?
It is likely you don’t have that data
Now you have a segment of traffic that are early adopters which doesn’t click on as many ads.
They may have switched off various forms of behavioural targeting so the ad payout on display advertising is less.
Early adpters and especially our SEO friends are notoriously difficult to convince to do anything that might make you money.
It seems to me that your conclusions are not valid. More likely than not, you still has the same type of visitors in the past, just not as a separate segment, so all data averaged out. Google’s providing the keyword or not has no impact on user behavior obviously.
Hello Andy,
Part of the issue is that you can’t really ever get that exact data prior to (not provided) being rolled out, unless you reprocessed your Server log files. It was more a view into we are going to get (not provided) traffic but will that traffic mask valuable keyword data that drives a large % of business or is it just tyre kickers.
But you are right some traffic is difficult to make any revenue off but that’s half the fun trying to crack the code to work out how you can make (not provided) traffic valuable.
[quote name=”Andy Beard”]Hi David
Possibly a few wrong conclusions.
Not Provided are early adopters
Were early adopters equal for conversion previously?
It is likely you don’t have that data
Now you have a segment of traffic that are early adopters which doesn’t click on as many ads.
They may have switched off various forms of behavioural targeting so the ad payout on display advertising is less.
Early adpters and especially our SEO friends are notoriously difficult to convince to do anything that might make you money.[/quote]
I think there is enough data I explored but you now have to think there is a new segment (not provided) that you could not previously split it. The segment is that and I see that you now have to consider and possibly plan for, some sites because it is a much larger % of total traffic. What type of content they consume, what changes or improvements could you make to improve the conversions/revenue from this segment.
I’m not saying it’s so much as different user behaviour it’s just a different segment, such as referral traffic, adwords…
[quote name=”Marcin”]It seems to me that your conclusions are not valid. More likely than not, you still has the same type of visitors in the past, just not as a separate segment, so all data averaged out. Google’s providing the keyword or not has no impact on user behavior obviously.[/quote]
David,
Thanks for a brilliant insight.
The fact that the (not provided) traffic converts at a noticeably lesser rate is not a surprise. Great to see some decent stats on it though.
I think Andy has hit the nail on the head with “early adopters”. I’d agree that these (not provided) visitors are savvy internet users who are permanently connected, rarely log out of any of their social networks, and suffer from terrible ad blindness.
The popularity of Google+ is still in its infancy so this trend is only set to continue.
What I think we will start looking at are ways of automatically tracking these visitors and serving them different content in the place of the traditional ads.
Tracking those visitors who are connected to OpenID is possible I’m sure, although we have not tested this out yet.
We will however have to do this at some point and our starting place will probably be here: http://stackoverflow.com/questions/3675571/openid-autologin-janrain-example-in-php
Again, great post David.
Cheers,
Duncan
Interesting to see these figures but are a few assumptions being made I think. A lot of questions but very few answers.
Its good to see someone looking into this and helping to get some positive takeaways 🙄
A very interesting article David. Makes me want to launch into a piece of qualitative research to learn more about the characteristics of logged in users and their propensity to convert in different contexts!