Adotas

Where media buyers start online

Featured Author

Gary Angel brings over twenty years of experience in decision support, CRM, and software development, to SEMphonic. Gary co-founded SEMphonic and is president and chief technology officer. He's responsible for leading SEMphonic's development of Web analytics and SEM decision making tools for web marketing professionals. In addition, he helps companies like WebMD, Intuit, American Express and Charles Schwab maximize their web channel marketing through intelligent use of Enterprise Web Analytics

More articles by Gary Angel





Features

Contemplating Conversion: Why the “Measure for Conversion” Mantra Isn’t Simplistic Science

Written on
Sep 1, 2006 
Author
Gary Angel  |
Share
Contemplating Conversion: Why the “Measure for Conversion” Mantra Isn’t Simplistic Science

Let’s take these in order, starting with the most obvious.

In a multi-channel business, your website is just one conversion point. Whether it’s the best conversion point or the worst is going to depend heavily on the nature of your customers, your relationships and your sales process. If you do a lot of affiliate marketing, and your affiliates sell your product for 20% less than you do on your website, then it’s simply unreasonable to measure your website by conversion effectiveness (unless, of course, you are measuring it by conversion to affiliates as well).

For almost any multi-channel business, it will be essential to come to some understanding of how important the Web is to supporting other channels. This is rarely a simple answer to come by, but once arrived at it can greatly shape your thinking about how to measure the Web. You may need to find website points that correlate to purchase interest and measure engagement to those points instead of measuring to actual conversion.

Our second point deals with the effect of visitors who aren’t coming to the website to purchase. If your website has a Customer Support component, for example, then it is simply wrong to measure your conversion rate as total conversions / total visitors. Including Customer Support visitors will obviously depress your conversion rates. That, in and of itself, isn’t important.

But what if your Customer Support traffic spikes every time you issue a new release? Now, your conversion rate is going to tank when you launch a new release — possibly setting off an entirely unwarranted firestorm of concern! The more functions your web site is being asked to perform, the less relevant conversion rates are as a measure and the more they will be distorted by outside factors if you try to trend them.

Fortunately, you can often segment visitors based on high-level cues from their behavior. So you can create a segment of visitors who view support information and subtract them from your total visitor counts. The more non-purchase segments you can identify and eliminate, the more meaningful and accurate your conversion rates are going to be.

If you can’t easily find behavioral cues that identify negative intent, you can also use measures of engagement to better classify potential prospects. It is sometimes easier to find pages that indicate a visitor is an interested prospect. By tracking conversion versus interest, you lose part of the funnel of your website (not all prospects may make it to the “interested” pages) but you have a measure that’s likely to be significantly more representative.

Our third bullet deals with the relationship between source and site efficiency. If you are trying to determine the effectiveness of your website or its “convincing” content, then you need to recognize that the level of qualification of prospects is going to be a significant factor in real-world performance.

Suppose you measure your website performance and find you have a 4% conversion rate. You then launch a big SEO program and your conversion rate falls to 3%. Has your site gotten worse? A significant percentage of new lands from SEO improvements or expansions in search or advertising programs will nearly always lead to lowering in the overall qualification of your prospects. The law of diminishing returns nearly always applies here. So site efficiency can only be trend measured vis-ÃÂ-vis a specific marketing mix. Failing to understand this can result in a series of totally unnecessary site-improvement “fire-drills.”

The fourth bullet point may be the most subtle and least welcome to web marketers. The idea is that you want to understand the operational conversion efficiency of your web site (its ability to convert those who decide to convert) and your marketing efficiency — its ability to convince real prospects to buy. A significant percentage of your web site conversion is quite likely to be purely operational — because the visitor showed up on your web site fully intending to buy. How can you know this. You can’t always.

But if your sales cycle is typically fairly long and the median conversion for visitors is 3 or more visits, then it’s a pretty good bet that any visitors who converted in their first session were fully intending to buy when they showed up. That means if you really want to measure the “convincing” performance of your site, you need to remove the visitors who’d already decided to buy just like you removed the visitors who aren’t prospects at all!

The last point addresses a common problem in almost all web measurement — the lack of statistically measurable significance of most changes on total site conversion. Most real site design decisions are at the page or even sub-page level. How should a message be worded? Does an image work better here? Is their too much text on this page? Even A/B tests to try an answer these questions typically won’t yield an answer when the outcome you’re measuring for is conversion. That’s especially true when conversions are a fairly low percentage of total volume.

There are two paths to tackling this issue: measuring to conversion proxies and measuring to Functional KPIs (Key Performance Indicators). The idea behind Functional KPIs is that you can measure micro-changes on a web site by using KPIs that are specific to the page’s actual function. In many cases, these KPIs let you measure the effect of very small changes to a single page.

None of the ideas described here really challenge the ultimate truth of the “measure for conversion” mantra. But together, they should make it plain that a simplistic view of what a conversion is and how conversion rate should be measured can be seriously wrong.





Reader Comments.

Gary,
Love the article, and appreciate the points you are making.

But you seem to be both rejecting and embracing the idea of the “site”. You clearly don’t believe that sitewide conversion is precise enough, but you don’t seem willing to embrace the idea that you should simply stop measuring the site and start measuring campaigns only.

Check out my blog, The Site is Dead. I try to address this issue from multiple angles.

I would also add that conversion rate is a lousy proxy because it does not contemplate lead value. For ecommerce, you really have to measure revenue (hard) and even for publishers and other web properties, you should establish value proxies for critical user events.

Matthew Roche
Offermatica

Posted by Matthew Roche | 11:07 am on September 6, 2006.

Gary: The point you make on why conversion rate is not great but then your proposal is still very fragile based on “behaviour identification” based on page views. This can be wrong for many reasons (think SEO directing people deeper into sites, where they don’t want to be, or wrong clicks becuase of bad navigation and many others).

Occam’s Razor blog by Avinash Kaushik had a very good post on “Stop Obsessing About Conversion Rate” which also proposes a much better metric to replace conversion rate:

http://www.kaushik.net/avinash/2006/07/stop-obsessing-about-conversion-rate.html

That blog also has a post on how to measure conversion rate for maximum impact:

Conversion Rate Basics & Best Practices

http://www.kaushik.net/avinash/2006/07/excellent-analytics-tip5-conversion-rate-basics-best-practices.html

-Peter.

Posted by Peter Cohen | 1:36 am on September 19, 2006.

Leave a Comment

Add a comment

Pages: « previous page 1 2

No Tags
Article Sponsor

More Features


  • Right now, at the beginning of 2012, what are you watching the most closely for its ad and marketing opportunities?

    View Results

    Loading ... Loading ...

Latest News

News Archive

Spotlight

Sponsormob Leads the Way Into RTB for MobileADOTAS – For more than half a decade, Berlin-based tech firm Sponsormob has remained relevant in an industry characterized by [...] more...


Adotas Partnership