Learn Methods, Metrics
(site map)

[ Home ]    [ FAQ ]    [ Download ]    [ Contact / About / Privacy ]

Get the Drilling Down Book!
http://www.booklocker.com/jimnovo

Prior Newsletters:
http://www.drilling-down.com/newsletters.htm
-------------------------------

Hi Folks, Jim Novo here.

This month, we're working through the measurement of Visitor Retention on the web.  I'm surprised at how many web analysts think Frequency, or Engagement if you like, is a measure of Retention.  Seems logical on the surface, but if you're only measuring the Engaged, how does that tell you anything about Retention?  What about the Visitors who were formally Engaged, the now dis-Engaged?

Over on the blog, we have an example from the Lab Store on economic decision making in a Relationship Marketing environment.  Risk versus reward is the issue, the question is, would you take the gamble?

Finally, I'd like to remind folks about the Predictive Analytics World conference going on in San Francisco Feb 18 - 19.  I can't make it myself, but for those of you ready to move beyond the kind of simple predictive modeling I talk about, it's simply the best place to be:

http://www.predictiveanalyticsworld.com/

Sample Marketing Productivity Blog Posts
==========================

Relationship Marketing Economics
January 9, 2009

This is an example from the Lab Store (my wife's e-commerce business) outlining the economics of decision making and profitability in a customer centric business.

Continue reading on the blog:
Relationship Marketing Economics
and feel free to leave comments.

Questions from Fellow Drillers
=====================

Visitor Retention Mapping

Q:  The research folks in my company are trying to convince me that measuring sessions and Page Views per Session is more effective than using Recency / Sessions, as advocated in your book, for a retention metric.

A: For a content site, the Page Views / Session measure can be used as a measure of visitor quality and appropriate marketing to the right audience - a customer acquisition idea - not retention.  And it really needs to be broken out by Source - the average has little actionable meaning.  You want to know the Visitor Sources, and then look at this metric by Source.  This is still Frequency though...what about visitors who don't come back?

Q:  I am having some difficulty in making a decision regarding this.  They want to give me a matrix with Page Views per Session on the Y axis and Total Sessions on the X axis to give me the "customer retention map".

A:  That's a Customer Activity map, I don't see how it has anything to do with Retention.  This is a "Frequency Map" with no Recency, which is more of a visitor quality thing, it doesn't tell you anything about Retention.  

For example, as long as you are growing your visitor base, with this kind of map you would never realize Visitors were defecting until the Visitor totals started to go down.  By then, it would be too late to take action, the damage is done and you're into the death spiral.

Amazon used to boast about how many "best customers" they had based on Frequency.  When Wall Street found out 60% of these best customers had not made a purchase in over a year (Recency), the stock was cut in half.  Frequency is a measure of Value, it doesn't mean anything for Retention without Recency.  Recent-Frequent visitors are best customers, now and in the future; Frequent visitors who are not Recent are former best customers.

Further, without Visitor Source, I don't know how you take action on the reporting above.

Q: They say that if I need to send messages to only the recent users I could by only querying the users with high sessions that were only generated over the recent past, say, 1 week.  This will be enough to target the users who you say are most likely to respond.

A: Well, this is true, they are saying to pull the list by Recency and Frequency.  So if targeting an e-mail to most likely to respond is what
you want to do, this would work.  But it doesn't give you a "Visitor Retention Map", if that is what you want.  It doesn't tell you how to get and keep best Visitors, or which Visitors seem to be defecting - implying weak Source.

And, response is not always what you want. More often, what you really want is increased profits.  So if you are churning through a ton of best customers each week, I would be more concerned about mailing to best customers who are not coming back.  You can't do that without Recency.  Sounds to me like they want to take the easy way out.  It's easy to pull a list of people who have been to the site, harder to pull a list of people who have not been there, which is what you want if you intend to act on Retention data.

Q: What are we missing here?  Why, if at all, is it more important to gauge Recency and Frequency (sessions) than Sessions and Page Views per Session?  It sounds like Page Views per Session is more like "Intensity" than anything else, and I can see how it is also valuable.  Can you shed some light here?

A: Yes, Intensity, which implies Quality, as I said earlier - for most sites a desirable thing. It's just another way to measure Frequency. So if this metric is more important than Total Page Views or Total Sessions to your business, then use it.  But without Recency, you have no predictive power.  If you are going to do list pulls based on Recency but not track based on Recency, it seems to me you will be missing the data you need to achieve your retention Objective.  How do you know if it's working?

Under this method, you could start with 1000 Recent Best visitors and work your way all the way down to 10 and never learn anything or come up with a plan to reduce defection.  Worse, this mass defection could be covered up by Acquisition efforts without the Recency  reporting, meaning you'd just be churning Visitors and not making any progress.

For example, say you are running two ad campaigns.  Ad A generates Views/Session of 6.4, Ad B Views/Session of 3.2.  Based on the tracking proposed, Ad A generates the "best customers".  So you put all your money into Ad A, and kill Ad B.

Later, you want to do a high-response e-mail, so you pull all the 1 week Recent visitors with high Views per Session, send it out, and get high response.  So far so good.  On further analysis, you find most of these people originally visited as a result of Ad B.  Curious, you look at the Recency of Ad A visitors and find 90% never came back after the 1st visit.

So you have just wasted a ton of money and time, because you were not tracking Customer Retention, just using a Retention-oriented metric it to pull lists.  See the difference?  The value of this kind of modeling is to be able to predict the future, not react to the past.  I don't know what you are trying to accomplish in the end, what is the Objective of all this?  But assuming high Page Views / Session is good, then you have a way to track Frequency, which is one vector on the Customer Retention Map, the Current Value. But what good is knowing "initial quality" if this quality doesn't stick over time?

That's the other half of the Customer Retention map, the Potential Value.  And that's where Recency comes into the picture.  If you intend to increase Retention, try to keep your customers longer, you want to:

1. Understand which Visitor sources create Visitors that do not come back, and if you are paying for any of these sources, reallocate the budget to those sources creating Visitors who do come back again and again.

2. Take action with the folks that were high value Visitors who have stopped Visiting, not the ones that still are.  Getting a high response rate from Current Visitors doesn't imply anything about Retention; it just tells you that engaged visitors are more likely to respond.

What you want to know is what Visitor dis-engagement looks like, how many high value Visitors are you losing?  Then find out why they are leaving and take action.

For more detailed examples, see the blog, Measuring Engagement Series.

Jim

Have a question on Customer Valuation, Retention, Loyalty, or Defection?  Go ahead and send it to me here.

-------------------------------
If you are a consultant, agency, or software developer with clients needing action-oriented customer intelligence or High ROI Customer
Marketing program designs, click here
-------------------------------

That's it for this month's edition of the Drilling Down newsletter.  If you like the newsletter, please forward it to a friend!  Subscription instructions are top and bottom of this page.

Any comments on the newsletter (it's too long, too short, topic suggestions, etc.) please send them right along to me, along with any other questions on customer Valuation, Retention, Loyalty, and Defection here.

'Til next time, keep Drilling Down!

- Jim Novo

Booklocker.com     Amazon.com     Barnes & Noble.com

Find Out Specifically What is in the Book

Learn Customer Marketing Models and Metrics (site article list)