Search Engine Marketing: How to Successfully Measure Your Web Site Analytics, Part 2

With the luxury of data in your hands, your ability to quantify your
recommendations will set your projects apart, and move them to the top
of the queue. This can be fairly simple. Whether focusing on leads,
purchases, call avoidance, click-throughs, downloads or any other type
of goal, give the achievement of the goal-or conversion-an assumed
value. For example, assume that a lead generated on your site is worth
$100 to your business and that there is a lead form that generates
50,000 visits per month. Over the last few months, conversion rates
have been relatively steady at about 18%. Given the assumed lead
value, the form is generating $900,000 per month. Through targeted
campaigns, navigation adjustments, form design changes and testing,
you estimate that conversion can be increased to as high as 23%. The
5% increase in conversions results in an additional $250,000 per month
or $3,000,000 per year. Now this project can be assigned a priority.
These recommendations should take precedence over any project not
anticipated to generate $250,000 per month. Put another way, every
month that the organization waits to implement the change costs
$250,000. Use a spreadsheet model so that numbers can be quickly
changed if assumptions are challenged.Use the Best Strategy:Assume
that you have run some stellar analysis and believe in ten
recommendations. You want to implement them all, but know there is no
way they will all be acted upon. Lead with the one that has the best
chance for success even if the quantifiable impact isn't as large that
of another recommendation. Once successful results have been realized,
this success will justify more initiatives. With each win, it will
become increasingly difficult for the IT organization to prioritize
other projects ahead of the ones driven by comprehensive web site
analysis.Lead with your best chance for success even if you can't
quantify the potential as being quite as large as one of your other
recommendations.Hypothesize and Test ResultsIt is imperative that
recommendations are positioned as hypotheses, and that all parties
understand that implementing a change hasn't proven anything. Assume
your company is running an email campaign designed to lure visitors
back to the site to buy laptops. The target population is 10,000
recipients. The test may be as simple as sending 1500 recipients an
email with subject line "Free Trial" and another 1500 recipients an
email with subject line "Limited Time Offer." If the "Limited Time
Offer" subject line results in a 10% higher conversion rate, consider
sending the remaining 7,000 recipients this version.