Web Measurement that Matters
It's been an interesting road over the past weeks. Now faced with mounds of data and the daunting job of learning new analytics tools. One good thing has occurred. In order to gain confidence in new business processes, I had to go back to class.
In my mind, there are several fundamental steps in ensuring effective Web Performance Management and here they are:
Ensure Tracking
Why would I include tracking? Its the nucleus of web analysis of course. The analyst relies on the efficacy of the data he observes through vendor tools. If the data is either erroneous or simply missing (perhaps undiscovered), he may make bad observations, or perhaps even more specifically, he misses opportunities; the fundamental objective of the analyst. Make sure your analytics tool can report on what you need. You may have to introduce a hack or two if needed, but get that data. Your conversion rates may rely upon it.
Further, poor observations can cost an organization thousands. Dare I say millions? No pressure. Many companies are willing to pay top dollar for blue chip analyts. They (companies) expect you to understand the nuances of effective vs. ineffective web tracking, independent of vendor selection.
Establish Benchmarks
No kidding on this one, right? Not so fast. When you've been analyzing and making decisions through analytics for awhile, you can to some degree predict success. Think about it. Ignore industry standards around email (2.5%) and direct mail conversion rates for a moment. What conversion rates are you used to seeing based on YOUR marketing efforts? Do your email conversion rates change depending on vendor? Offer? Audience? You say you aren't sure? Maybe its time to collect that information along with clickthrough and open rates? There are other metrics that take place offline that should be considered. Be sure to include them.
Set Targets
Another no brainer, right? How many organizations really set goals for themselves before they spend marketing dollars? Its irresponsible marketing, but we've heard it before. Start with the end goal in mind and work backwards. You have the benchmarks now set the targets. That may help you make better sense out of your marketing budget and better predict success. Is that Cost Per Lead allowing you to reach the black based on your spend or conversion ratios? Be sure to subtract marketing cost from your revenue goals to realize net profit. Finally, use your analytics tool to track ongoing success. Are you on target? Line that data up and then set alerts.
Identify the Metrics
At this point, most analysts should be familiar with identifying core metrics and setting KPIs based on the type of business model your site embraces. There are too many resources to illustrate so I won't cover them here.
Don't be afraid to create your own calculated metrics to help business owners better understand performance. At the same time, don't try to measure too much. More metrics doesn't necessarily mean better management. One final note, use the language that matters to the business. Don't call something a conversion if the business wants to call it a prospect. Don't call it a lead, if its really just a record. The language of online conversion is meaningless if it has no connection to the business.
Develop effective KPIs
I offer only one statement here. Keep it simple. Identify the levers to success. More often than not, measuring those things that affect money is a far far better use of the analyst's time.
Segmentation
There are many sides to the problem of analysis. The analyst should be able to identify top performing paths, referring URLs, keywords, affiliates, lead source, etc and tie that to online success or failure (mainly money). The analytic tool of choice should be able to help the analyst segment data/groups quickly (no 24 hour delays..apologies to some vendors). Segmentation is discovery. By its very nature it is often adhoc. If you want to manage your site's performance, you'll need basic level segmentation tools. No need to pick of that SPSS package just yet.
Share Results
Analytics is still so very new. To reference the title of Eric Petersen's book; 'Demystify Analytics'. Share data. Keep it in the context of business results and review marketing performance often. Set targets and reminders. Above all else, the analyst must share results and finally, make appropriate changes that make the metrics matter to the business.
In my mind, there are several fundamental steps in ensuring effective Web Performance Management and here they are:
Ensure Tracking
Why would I include tracking? Its the nucleus of web analysis of course. The analyst relies on the efficacy of the data he observes through vendor tools. If the data is either erroneous or simply missing (perhaps undiscovered), he may make bad observations, or perhaps even more specifically, he misses opportunities; the fundamental objective of the analyst. Make sure your analytics tool can report on what you need. You may have to introduce a hack or two if needed, but get that data. Your conversion rates may rely upon it.
Further, poor observations can cost an organization thousands. Dare I say millions? No pressure. Many companies are willing to pay top dollar for blue chip analyts. They (companies) expect you to understand the nuances of effective vs. ineffective web tracking, independent of vendor selection.
Establish Benchmarks
No kidding on this one, right? Not so fast. When you've been analyzing and making decisions through analytics for awhile, you can to some degree predict success. Think about it. Ignore industry standards around email (2.5%) and direct mail conversion rates for a moment. What conversion rates are you used to seeing based on YOUR marketing efforts? Do your email conversion rates change depending on vendor? Offer? Audience? You say you aren't sure? Maybe its time to collect that information along with clickthrough and open rates? There are other metrics that take place offline that should be considered. Be sure to include them.
Set Targets
Another no brainer, right? How many organizations really set goals for themselves before they spend marketing dollars? Its irresponsible marketing, but we've heard it before. Start with the end goal in mind and work backwards. You have the benchmarks now set the targets. That may help you make better sense out of your marketing budget and better predict success. Is that Cost Per Lead allowing you to reach the black based on your spend or conversion ratios? Be sure to subtract marketing cost from your revenue goals to realize net profit. Finally, use your analytics tool to track ongoing success. Are you on target? Line that data up and then set alerts.
Identify the Metrics
At this point, most analysts should be familiar with identifying core metrics and setting KPIs based on the type of business model your site embraces. There are too many resources to illustrate so I won't cover them here.
Don't be afraid to create your own calculated metrics to help business owners better understand performance. At the same time, don't try to measure too much. More metrics doesn't necessarily mean better management. One final note, use the language that matters to the business. Don't call something a conversion if the business wants to call it a prospect. Don't call it a lead, if its really just a record. The language of online conversion is meaningless if it has no connection to the business.
Develop effective KPIs
I offer only one statement here. Keep it simple. Identify the levers to success. More often than not, measuring those things that affect money is a far far better use of the analyst's time.
Segmentation
There are many sides to the problem of analysis. The analyst should be able to identify top performing paths, referring URLs, keywords, affiliates, lead source, etc and tie that to online success or failure (mainly money). The analytic tool of choice should be able to help the analyst segment data/groups quickly (no 24 hour delays..apologies to some vendors). Segmentation is discovery. By its very nature it is often adhoc. If you want to manage your site's performance, you'll need basic level segmentation tools. No need to pick of that SPSS package just yet.
Share Results
Analytics is still so very new. To reference the title of Eric Petersen's book; 'Demystify Analytics'. Share data. Keep it in the context of business results and review marketing performance often. Set targets and reminders. Above all else, the analyst must share results and finally, make appropriate changes that make the metrics matter to the business.
Labels: site management
<< Home