Chinalytics Longmarch Blog

In a previous post I expanded on my methodology to reach actionable insights through web analytics by answering the 4 questions:

1. What is the current performance?





2. Is it good or bad?





3. What are the root cause and key drivers that lead to the current performance?





4. Which steps can be taken to improve the current performance?

While discussing the importance of context to answer question 2, I did not touch one key point:

• Did we reach our target?

Setting and measuring on targets is critical to all analytics efforts since, as Seneca said in my post title, "If a man does not know to what port he is sailing, no wind is favorable". Yet often targets for web analytics remain elusive.

Online marketing and campaign objectives are often vague and unspecific. Take the objective of this blog for example: "Provide relevant web analytics insights and best practices to local practitioners and drive the adoption of an analytic mindset for online marketing in China". All the detailed numbers Google Analytics and Feedburner can provide cannot tell me if I reached this objective.

To make the objective measurable, we need to break it down into measurable Key performance indicators (KPI) and to measure success we need to set target values for these KPI’s. If that sounds more like business school gibberish than down to earth web analytics (?) remember the old adage "You get what you measure". So lets break the objective down into pieces to see which KPI’s we can identify.

relevant content:Relevance is always in the eye of the beholder, but I can track the number visitors who identify this blog as relevant by tracking their loyalty e.g. repeat visits and subscription to the RSS feeds.
local practitioners: Since this blog is targeted at Chinese (what a shame I can’t write Chinese), the KPI should be Visits from China / Visits.
Adoption of analytic mindset: While I can’t measure what is on peoples mind (though not for the lack of trying), I can measure actions. In order to have an impact on the ecosystem, I need to reach a large number of people and engage them in a discussion about analytics. Both are measurable. For reach I will track Unique visitors. And for engagement tracking comments, trackbacks & emails received is relevant.

Now that we have a spanking new set of KPI’s, we need to take the next step and set targets for each KPI.  Many web campaign and website owners do have a set of KPI’s (not necessarily the right one), but don’t have a target. The analytics team then end up in a desperate search for context and comparables to tell them is they are doing well. There is no context like a wisely set target, since the target setting process should take all available context into account.

For my little red blog I set the targets for October like this

Reach: Unique Visitors: 200
Loyalty:  Repeat Visits + Feed subscriptions: 100
Engagement: Comments + Trackbacks + emails received: 15
Visits from China: 70%

Since I am new at blogging, I am not sure how realistic these goals are. But I will certainly report my progress. If you have any thought, comments or recommendations. Please share them with me and my readers in the comments.

Microsoft aQuantive acquisition and Google’s recent move to hire Madison Avenue talent has created a lot of controversy about their future plans and how traditional agencies should respond. Some argue this heralds the end of the agencies and all of us need to pack their bags.  While I don’t agree with this doomsday (at for agency people like myself) scenario, it is certainly interesting to hear what Google and Microsoft have to say about the future of online marketing.
Rebecca Lieb of ClickZ, one of my regular reads (see blogroll), shared some very insightful thoughts on this topic today. The two key points that make me think are:

1. "You don’t need to go into a fixed budgetary cycle," Armstrong argues.
   Instead, advertisers must learn how their budgets fluctuate
   with consumer-interest cycles.





2. Utten may have put it best when he explained insofar as E-Trade’s
   constellation of traditional, interactive, and media agencies are
   concerned, the company remains idea agnostic. Instead, what he’s buying
   is the "best optimization."

While idea 2 is not really new, and everyone in web analytics space should live and breath it already, the first idea takes online advertising and web analytics to an new level. It really is the next big step away from the broadcasting marketing model of old to a targeted, focussed and optimized advertising model. It recommends moving away from spending money communicating to your customers when you want to say something, to spending money communicating to your customers when they want to hear from you. That for once sounds like an idea that moves the needle on marketing effectiveness. While I continue to wrap my head around this, I hope to hear your thoughts in the comments section.

If you have not read Avniash’s Book (and you should), you have to read and watch this post. Its a great overview of his ideas, and provides a great framework on how to think about web analytics. What, you still haven’t clicked?

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.

In my last post I mentioned that we are currently using GA Analytics for one of our projects. So far I have to say, it is one sweet tool. It is easy to implement, has a very intuitive, easy to use interface and all the key metrics are there.

Some of my fellow analytics bloggers seems to believe that making such a great tool available free of charge is Google’s bid to monopolize the Web Analytics Tool market place (ala Microsoft and Internet Explorer). I believe nothing could be farther from the truth. In fact, I am sure Google loves any kind of Web Analytics Tool, from Omniture to Webtrends, from Clicktracks even to Microsoft’s Gatineau. That is because any Web Analytics tool, used well, will tell you that Search is the most effective driver of high quality traffic. And search, after all, is where Google is raking in the money.

Now, how to let EVERYBODY know that? You can tell them, again and again (and again). Some may believe you, other will think it’s just a sales pitch. So how about giving them the tools, to find out themselves and improve their whole online marketing process along the way? And giving them these tools for FREE?? This is what Google Analytics is all about.

But here is the problem: Now everyone has a great free tool on their hands. But do marketers in China don’t know how to use it effectively? The answer is in China is certainly no! And there is noone around in China to tell them! (especially in Chinese) You will find a couple of Chinese Blog telling you how to use Google Analytics. But there is noone telling you how to do comprehensive web analytics. The kind that gives you actionable business insights. That is what Google (Baidu) need an Analytics Evangelist for. Any takers for the job?

As one of the more regular topics on this blog will be highlighting some key metrics that every web advertising campaign and web site owner needs to pay attention to, and how they relate to online campaigns in China in particular. These will go in no particular order, but I plan to cover the ones I see as most critical for China first.
The first candidate is the "Bounce Rate". The Web Analytics Association defines Bounce Rate as Single page view visits / Entry Pages . A more common definition used by most analytics tools is: Single Page Visits / Entry Pages. It measures how many of the visits that used a page as the entry page to your site, were single page visits. Or in marketers terms: Is your landing page interesting enough for users to take action (clicking links on the page).
Now why is this so important for online campaigns? In traditional media advertising, we analyze TV ratings, media circulation and other metrics to understand how many people have been exposed to our AD. Then we do surveys to understand if the AD broke though the media noise and how it affected the audience’s attitude.  Most of the time, this is exactly what we advertisers in China are doing for online advertising campaigns as well. We measure the impressions recorded on the page the AD was displayed, then measure the clicks and click through rates on the banner (or search keyword in too few cases) to measure the "break through" of the AD, and then we survey the audience about brand preference etc. What is often ("most of the time" in China) completely ignored, is what happens after the users clicked on the banner! Since this is what makes web advertising so unique and in my opinion uniquely valuable, this is a complete and utter waste. Once a visitor clicked on our ad, we know that he is interested in our message. Even better, after the click he comes to an environment that we have 100% control of. This is the place where we can really impact the visitors/users/potential clients perception of our brand, increase his brand awareness, increase his brand preference of even let them buy our product / service directly.
The bounce rate tells you how well you have made use of that opportunity. How well you have used the web to do more that just show users pretty ads through a new channel. The answer is, that in China, at the moment, sadly, we are doing very badly. My discussion with clients, friend and industry experts suggest that bounce rates for online campaigns in China hover between 90 and 99% percent, with good campaigns coming closer to 90 but only rarely below. Comparisons show that the average bounce rate is closer to 60% in the US (in lieu of other data points). It means that, of the 6-7 billion RMB spend on online advertising in China in 2007, more than 5.5 billion fail to use the online channel effectively. In other words: If we come close to the US numbers, we can improve our online advertising effectiveness but at least 300%.
This is certainly a number to think about. Even though there are several factor that might help to explain parts of this outrageous number,

• Many netizens in China might not be able to differentiate well between content they are interested in or all relevant content is already available in. 





• the landing page might contain all the info the visitor is interested in, and he/she does not need to click further





• Sometimes visit where the visitor only clicked on a download link counted as bounces

it is critical for every advertiser to understand what their bounce rate is, how it relates to their campaign goals and to the bounce rate of competitor’s campaigns. So here is a question to my readers (the one or two who might read this), what is your bounce rate, and what do you go about reducing it?

Shortly after I started my new career in Web Analytics, my friend T.R. Harrington from Darwin Marketing asked me which analytics tool I could recommend for companies in China. I could not answer his question then, and it is even harder to answer now.
During the last months I realized, that, although China has the second largest internet population in the world, when it comes to analytics and optimization it is still in its infancy. I attribute this gap to 2 key factors:

The lack of a mature e-Commerce environment: E-commerce sites earn the most immediate benefits from detailed web analytics. They are usually analytics pioneers since effective web analytics directly benefits their revenue.

The limited understanding of web marketing on the side of major advertisers, which are not aware of the
opportunities for “performance based marketing” the web offers, and how analytics  opportunities exceed what is possible in traditional TV, newspaper &
outdoor advertising.

Luckily at least the second factor is slowly changing, which is why my role exists and why Baidu’s stock price is continuing to rise. But since the e-commerce pioneers did not prepare the field for us, many basic tools and lessons are still missing.

I am not aware of any major analytics vendor (like Omniture or WebTrends) that has servers in China. Which is important since the slow intercontinental connections and Great Firewall can impact data tracking. Having used both packages, it becomes obvious that they are lacking in other areas as well. Omniture for example cannot correctly read search terms typed into Baidu, while WebTrends does not allow a regional analysis of Chinese traffic (and lists Chinese cities that no one in China has heard of). We will be running a trial with Google Analytics in the next weeks and I will let you know my impressions. But until then, please let me know your experience with Analytics tools in China in the comments section.

After joining the web analytics industry a couple of month ago, I found a wealth of information online. Especially Avinash Kausnik’s Blog and book were a great gateway to the world of web analytics. What became apparent quickly however, is that there is almost no, especially English, information available on “Web Analytics with Chinese characteristics”. This is a niche I will humbly trying to fill, while helping my clients along the Long March towards accountable, effective, and efficient online marketing.