Marin Software's official blog

Defining an Analytics and Optimization Plan

|
August 19, 2011

Analytics is the process of using historic data as the basis of any new decisions. Online advertisers now have an increasing amount of tools bringing ever-improving reporting functionality. The number of metrics being recorded and the volume of data being captured leads to the question - where should analytics and optimization efforts be focused?

Apart from the volume of data, online campaigns are especially difficult to analyze since there are so many factors constantly changing. The performance is directly affected by the actions of competitors. Also, there are often multiple people working on the same campaigns each optimizing to their own agenda, but in doing so, making any analysis of overall performance far less clear.



As mentioned, there are numerous metrics for analysis and optimization, some of the key areas that affect online campaigns are:

  • Coverage
  • Creative messaging
  • Seasonality
  • Day-parting
  • Ad position
  • Bidding strategy
  • Paid vs. Natural traffic
  • Cross-channel attribution
  • Landing page
  • Bounce rate
  • Time on site
  • Conversion journey


The importance of each metric will differ for each client, but narrowing down your options to what is currently possible is a great first step in building out your plan. For example, if you are unable to influence the journey from landing page to conversions, then you should probably disregard this in favor of something else! Similarly, if the tracking solution currently employed only tracks conversions to the last click, attribution analysis will not be possible. Equally, if the recent period is unusual, such as a sale or seasonal uplift, some analysis might best be delayed till later in the year.

Working on all of these areas at once will make it very difficult to determine which changes are generating the improvements, and just as importantly, which are not proving to be a good use of your time. A clear strategy should be worked out based on the areas would yield the highest potential improvement in campaign performance. For example:

  • if the CTR is low, work on your ad copy/negatives and targeting
  • if your conversion rate is low, optimize your landing page and path to conversion
  • if spend volumes are low, then work on both of the above and your bidding strategy
  • if conversions seem to be driven solely by brand terms, then look beyond the last-click to investigate what is driving these brand searches and clicks?


During a period where a particular aspect is being worked on, all changes should be tagged so that it is easy to filter and analyze the performance before and after testing to draw conclusions as its success. Ensure you are using a tool that allows you to easily tag changes that you are making and make notes for all users so that any sudden increases/decreases in performance are noted in a single area. This will be hugely beneficial when making decisions weeks or even months later.

Sign Up and Get Updates by Email
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.