Thursday, December 27, 2007

Web Analytics powered 1to1 targeting

The First Step Change (As I had mentioned in my previous blog about the 3 step changes in Web analytics) in my view is how online marketing (Banner advertising, PPC, affiliates, email and other online marketing channels) and behaviour targeting are putting to effective use the power of Web Analytics.

Typically there has been a big gap in spending when it comes to off-site marketing channels like

- Email marketing
- Affiliate programs
- Search Engine Optimization
- Offline marketing to web
- In-store Web promotion
- Banner advertising

And On-Site resources like

- Optimization
- Web analytics
- Usability testing, etc.

With the former hogging the major share and the later (on-site resources) taking a back seat most often.

However more and more marketers are realizing that though the off-site marketing activity drives all the traffic to the website. It is the landing page/ homepage or the other critical product pages determine if the prospect converts to a customer or not and how much he / she would spend on the site products or services.

Right after the prospect lands on any of the pages in the site and interacts on the same determines his/ her engagement with the site.

You have probably experienced at some point that you were searching for a product that you are interested in, clicked on a seemingly relevant search result and landed on a page which has no flavour of the product that you were searching for. Could you recall your reaction…it most probably was, Uhh…

This is what the marketers are striving to eradicate and make each visit to the site by the prospect as engaging as possible. This could be achieved by accurate one to one targeting.

One to one targeting or behavioral targeting is achieved by the following process:

a) The visitor arrives at the website.
b) The information goes to the “Visitor profile repository, to figure out if he is a new customer or an existing customer.
c) If he is a new customer, we will build the customer profile.
d) If he is returning customer then we retrieve his profile.
e) Ping the self learning predictive modeling engine determine the suitable or optimal content that we need to show
f) Pass this on to the CMS system, which will then serve the relevant page to the customer.

Apart from the technical solution of how to serve the relevant pages to the visitor, there is a bigger task of determining the criteria (to be fed into the predictive modeling engine) of what pages/ information to serve to the relevant customer.

If we could answer a few basic questions, we would be able to target our highly predictive anonymous visitor.

- What is this visitor doing now?
- What have they done before?
- When is this visit occurring?
- How frequently & recently have they visited?
- Where is this visitor Located?
- What is their online experience?
- How did this visitor arrive here?
- And Have they already expressed what they want?

We already collect all the data that we need to serve the relevant page to the customer

With the Site behavior variables we know the customer’s

- Previous visit pattern
- Previous Product interests
- Whether he is a New/return visitor, etc.
- Previous online purchases
- Previous Campaign exposure
- Previous campaign responses

We collect temporal variables like the time of the day, day of the week, Recency & frequency

We get Environmental variables like
IP address, country, city, browser type used, etc…

We also get the referrer variables like

- search keywords
- affiliate site
- campaign ID
- Direct/ bookmark
- Referring domain, etc..

With these details and an analysis of what is best suited to your site, the landing page relevance could be built and conversion rate, customer engagement with the site and loyalty could be enhanced.

This blog is inspired and derived from the presentation published by the eminent “Brent Hiegglke” the link to his presentation is as below:

Brent Hiegglke presentation

Behavioral targeting is gaining prominence because of the need of the marketeers to maximise the bang for the buck and also to ensure customer satisfaction & their loyalty. In the US alone the market is estimated to be 3,800 million by 2011.

In one example Brent had demonstrated how a leading European bank had adopted one to one targeting which was supported by high end analytics to drive relevance to the visitor. This Bank earlier had a generic landing page that displayed all its products and services based on some analytics such as conversion rate, etc. However when 1to1 targeting was launched it had a dynamic landing page based on the search question that the visitor had chose to land on the site. If the visitor was searching for loans, the leading banner would be their loan products banner, and if the visitor was searching for credit cards the lead banner could be the one with the same and so forth. This way the bounce rate from the landing page was reduced and a tremendous increase in conversion as well.

Also usually Behavior or 1to1 targeting is misunderstood as targeting to customers who come from another site reference or through marketing. It is also important to target customers based on their on site behaviour and their personal details. If your target audience is a working mother with young kids you would’nt want to show her a Ducati unless she has been viewing or searching for one. Similarly if a visitor spending a great deal of time and had numerous visits to the “High end plasma TV” should’nt we show our latest discount offer on the “Plasma TV’ or other high end entertainment devices in his next visit.

1 comment:

  1. This is a good informative articles. Thanks for sharing your knowledge with rest of the World.


Welcome your comments/ feedback: