-Avinash.

The importance of ‘conversion rate’ as a core analytics metric is being debated in blog posts from Avinash Kaushik, Matt Jacobs, and LunaMetrics. Avinash makes the point that conversion rate as generally defined is a vastly over-rated metric becaus…

http://digitalmediaanalytics.com/blog/2006/07/31/why-youd-better-obsess-beyond-conversion-rate/

What I love about this is that it expands the “conversion” concept to include people like learners, who are in the research stage of a process. However, I guess my concern would be that if you can’t tie the happiness of the learners back to the Bottom Line (ie, earnings for your business) in more than just a theoretical way, then it’s going to be hard to use these metrics to justify spending or make budget-based decisions.

This may also be a case of people being a bit afraid of knowing the truth about their website- ie, that a large percent of the visitors may *never* be paying customers!

Gradiva
http://www.yourseoplan.com/

But for many websites the tail is really long but in reality around 10 to 15% of keywords represent a majority of the traffic, in which case your idea will work perfect. The challenge, IMHO, is to find enough keywords in our top 10% where we can attribute intent because mostly traffic is driven by generic brand and category terms (which makes is harder to attribute intent).

You could go one better, if you have a truly dynamic platform, say like we have with ATG, it can detect that you came on “buy Canon sku xyz” and the platform can change content on the page to be a bit more targeted to what you need. So you don’t run around creating landing pages (and I am not implying that you do :)), the page stays the same, content changes on detection of intent as defined by you for keywords.

Thanks for commenting, that was a great suggestion.

I apologize for not completely understanding the concept of conversion between steps of the funnel and being unable to reply.

But your thought on measuring different stages, lifecycle, is absolutely the right thing to do. I was in a presentation at another company yesterday and saw them use the lifecycle: Learn, Shop, Purchase, Support. Very focussed, targetted and helpful.

My caution would be not to infer any kind of “conversion” from a page view. So “you come to my site, we detect you don’t have a cookie, you saw the page for Gateway NX100X, you exit and so you are in the intent or consideration bucket.”

Potentially that would be sub optimal. There are alternatives to measuring if people who are truly “consideration” are moving along the life cycle and I would recommend some non-web analytics options.

Hope this is helpful. Thanks so much for taking the time to comment, it is clear you are doing some great stuff at Gateway.

The problem, as I see it, is to scrape all of your query logs and segment the search terms by the categories that are meaningful to your business. You might actually have to go several thousand terms deep (ie: in Surfaid - now CoreMetrics) you can go down 50,000 deep for a background report. If there was a way to write a program that would take your query log and segment it into say…..LEARN, SHOP, or BUY….and then tie it to the pages the searcher landed on …. you’d have something very intersting to look at.

That would be a major program - and has yet to be written.

Avinash, would you say that it would be a valid approach to try and understand visitor intent by looking at referring keywords? I.e. perhaps have some kind of “bucketing” mechanism whereby you would group such keywords according to visitor intent? So, keyword “buy Canon sku xyz” might go in the “Sales prospect bucket” whereas “review Canon” might go in the “Browse/research” bucket. And once you have those buckets in place you can then start addressing the needs of those different types of visitors.

Cheers,
Michael Whitaker

My goal is to try to measure conversion along the funnel between each step of the customer lifecycle: unaided awareness, awareness, positive perceptions, consideration, intent, customer, satisfied customer, loyal customer, and advocate.

Of course, these can be combined into fewer steps, and measurement is always problematic, but just the act of thinking about the customer’s lifecycle in these terms has been helpful to us (even when we violate what you and others said at eMetrics — sometimes we measure things even though we can’t make decisions based on them, d’oh!).

What do you think?

Wandering Dave

Excellent post and a lot to think about here; I also read the comments which were also very helpful.

I have found that a particular metric - conversions / task completions are best defined in partership with the site owners - just as keywords for your pages need to be defined by the site owners/stakeholders in partership with your SEO/SEM person.

The ability to segment your visitors in ways that are meaningful to your site is probably the most important step prior to measuring the conversion (ie: segmenting by intent or segmenting by industry, etc). If we can figure out who the visitor is and what segment they belong to - then what we’ll call the conversion or task completion comes into view.

I’ll see if I can write a post on this tonight on Webmetricsguru.com; pulling in my own experience and ideas. One thing I’d like to see dealt with is the effects of Multi Channel Conversions (as I wrote about in http://www.webmetricsguru.com/2006/07/measuring_multi_channel_market.html ).

I’d be curious to see what weight is put on the various points that lead to a conversion - in this case awareness that leads to seeing a movie.

Measuring Revenue Per Visit is sub-optimal when a minority (less than 50%) of site traffic comes to buy. We are inflating the denominator and making ourselves look worse.

This might be better: Revenue Per Visit for Visitors who came with the Remotest Intention to Buy. That would be cool, but nearly impossible without chips implanted in our visitors brains that communicate to us their intentions.

My advocacy is for a step away from the whole “revenue” mindset to a “task completion rate” mindset.

For customer segments who come to buy, or with any intention to buy, measure conversion rate and revenue and all that good stuff (task completion rate will give you a similar feel for segments who are there to buy).

But for other customer segments don’t measure any revenue metric because that is not why they are there, try Task Completion Rate. Then we solve holistically and for the long term.

Does this make any sense? : )

Thanks for commenting Eric, it is so great to hear from readers of the blog.

As always, your devoted reader.

Eric T. Peterson
http://www.webanalyticsdemystified.com/weblog

Most often though conversion rate is connected to selling or submitting a lead etc. Hence my recommendation to use Task Completion Rate as the new metric. It could be “rose by another name” but in my mind it is a different metric that focuses on getting people to complete the tasks they came to the site for.

I am pessimistic that if we keep the metric the same, conversion rate, and now say we are focused on helping all customer segments complete their task on our website for all the main purposes for which people come to the site for.

Staying with Conversion Rate might still promote the short term focused behaviour from company employees. What do you think?

Thanks so much Steve for taking the time to comment.

Robbin: I want to stress that I am not recommending measuring Conversion Rate by needs. Please see my reply above to Steve.

(This is because we all crawl and then walk and then run.)

I am being bit of a smart ass here but: In most Central Asian countries they never had land line telephones, they went from Morse code to cell phones and progress came faster, cheaper and more efficiently.

In my mind if there are people who are new and don’t care as much about conversion rate I would much rather they got brainwashed into mindsets like Persuasion Architecture to learn about their customer needs and then focused on measuring what the Task Completion Rate is by Primary Purpose on their website. They will make much more money.

Here perhaps I am being too optimistic. What do you think?

Thanks so much for taking the time to comment.

PS: Robbin I would caution against measuring “need” based on page views. All web analytics applications would be sub optimal at that.

(My next post, to reply to your last thought, is on basics and best practices for measuring conversion rate, because I don’t want people to think conversion rate is not important, it is but just not as much as all the hoopla that exists. I finished drafting it last night, would love to know what you think of it.)