Web Analytics in China

CTR and CPC Benchmarks for China (part II)

After getting all worked up about JPMorgan's metrics, my  colleague Sydney restored me to sanity today by pointing out that the table I was referring to, most probably was reporting only to Search advertising data. Can anyone confirm that the numbers make sense for Search? T.R.? Is there similar benchmarks for display advertising in the report? Too bad I was late downloading the whole thing as Kaiser Kuo recommended. But then again I was relaxing my mind and stomach from a stressful Germany holiday ;)

CTR and CPC Benchmarks for China

Thanks Kaiser for pointing out this article at the China Web 2.0 Review with a summary of the China section of the new JPMorgan “Nothing but Net” report from Morgan’s Internet analyst Imran Khan and his team.
Apart from the interesting market size predictions (past data curtesy od CNNIC) the most interesting info for me was the average CTR and CPC numbers supplied for the China market (curiously called Price per click). I have been looking for industry benchmarks for ages, and this could fill this gap. (Insert exclamations of happiness, for frenzy forward to friends here) But looking at these numbers, they seem to make little sense. Click-trough-rates I have seen globally (incl. China) are usually less than 1% (not >20%) and CPC in China varies between USD 0.02 to USD 1. RMB 0.40 seems out of this world. I checked with CNNIC and they claim innocence. Can anybody with a hotline to JPMorgan clarify?

Unique Visitors & the importance of context

One of the key metrics in every analytics report is "Unique (weekly, monthly, etc) Visitors". In its easiest form, it will tell you how many people visited your page or site in a given period. Even given the problems of accurately measuring the advertised number (done with persistent cookies), unique visitors is the best headline number to evaluate the success of a campaign and add significant value compared to looking at page views or visits. The key reason is that it can answer the key marketing question: "How many people (prospective clients) have seen my message (e.g. website).
But just looking at the plain visitor number does not allow us provide actionable suggestions. And this is what web analytics is all about. Even adding general ratio like Visits / Unique Vistor or Page Views / Unique Visitor do not provide much additional room for analysis.
There are three simple questions that help us on our path to actionable insights.

1. Is this number good or bad?
2. Why is the numbers good or bad?
3. What can we do to improve this metric and is it worth the effort?

Today I will focus on point one.

"Is this number good or bad?".

The answer this question, we need a little help from a friend. This friend is context. In order to understand if we are doing well or not, we need to compare our number to other numbers. Let me use the example of school boy that comes home from school with a "B" in Chinese. (Luckily he is preparing from the future) His parents will naturally look at the context to interpret his performance. They would ask themselves question like:

• What were his previous grades? (A "B" is great if he got all "C's" before, but a little disappointing when we was all "A's"
• How well were his classmates doing? (If it was a difficult test, and all other classmates got "C's", this B is great!)

We need to do same in Web analytics, when looking at numbers like "Unique visitors".

1. How did the number of unique visitors develop over the last month? Is it growing or declining? How fast?
2. How are your competitors doing in terms of Unique Visitors? Who is getting more visitors, who less?
3. What is your share of the overall internet population? (Here I need to leave my school boy behind) Is  Is it growing or declining?

Most practitioner of web analytics will easily do point 1. Heck, most analytics tools will do it for us. Point 2 and 3 are more problematic, especially in China.

Competitive information
... is hard to come by. In the US and Europe Comscore and Hitwise will provide such competitive intelligence. In China we are less blessed. Startup companies still rely on their Alexa ranking (if favorable), but Alexa numbers have been discredited for some time. The only other provider I am aware of that can provide competitive web information is IResearch. The company started to run a panel of about 50K internet users, and provides its client access to their surfing behavior (adjusted to represent the Chinese Internet audience). Due to my past occupation, I am not a big fan of IResearch, but after initial tests their numbers seem fairly reliable.
Based in this competitive information your are now able to tell you client (internal or external). "We have 1000 unique visitors this month. This makes us number 3 in the industry. The best competitor is X." for more details on competitive analysis, that a look at Avinash's post here.

Share of internet population (Reach)
Another way of providing context to your Unique Visitor number is comparing it to the overall internet population. This is especially critical in a market developing as quickly as China. CNNIC provides a semi-annual report on the size of the Chinese Internet population. If your number of unique visitors is growing by 10% annually, while the internet population grows by 30%, you certainly have a trigger for a more in depth analysis of the cause (Why in my previous list). If you need reach data for your site on a monthly basis, you can refer to IResearch as well, which will have a monthly reach for your site.

The importance of context is true for any web metric, and I will continue to highlight it future post. If you have any thoughts or comments, as well as any recommendation for better competitive information in China. Please share with us in the comments section.