Amadesa

This is post #2 in a series of posts that will discuss the optimization process from planning through analysis. If you missed post #1, please take a quick read here.

We have a good understanding where to start testing and with how many test versions to run for the traffic available:

• Highly trafficked pages that contribute to business objectives
• Pages that have 'improvable' KPI’s (high exit rate, low add to cart rate, low click through rate, improvable revenue per visitor, etc.)
• Pages / flows that have diverse; ‘segmentable’ traffic that may be influenced by segment testing and targeting

A key next step in the preparation process is to take a ‘traffic inventory’ of all pages that will be included in your testing. Traffic inventory; why is this important?

If you are an internet retailer driving revenue or customer acquisition, chances are you have a digital (and non digital) marketing mix assisting in converting users and driving revenues.

You may have:
• Email campaigns for awareness, specials, cart abandonment, order reminders, etc.
• Affiliate programs with different promotions, creative treatments and destination URL’s
• Search engine marketing campaigns (for all major search engines; PPC and SEO) for brand awareness, sales, lead gen and more
• Other tests running like live chat, ratings and reviews, product recommendations, infrastructure monitoring (Gomez, Keynote) etc.

Considering that you may have any number of the above programs running on a given day, week or month all with different messages, it’s important to take traffic inventory prior to test launch. All pre-click marketing activities affect post click behavior; it’s important to realize that testing will also be affected.

Example:
Say your search engine marketing team is running a Google PPC program similar to the following ad:

Due to the content contained in the ad, the potential customer is expecting free shipping on their order.

If a cross site optimization program is running that is testing the free shipping banner (hide, replace free shipping with free gift or coupon, or 20% off your next order), the free shipping promotion running on Google may be adversely impacted. Customers that click on the PPC ad are expecting free shipping.

Similarly, if your site receives a large percentage of traffic from email marketing, it’s important to know when, how often, and the content contained within each email program.

Below are some tips for dealing with multiple traffic sources within your optimizations:

• Take traffic inventory for each of your marketing tactics that may affect website traffic, conversion rates, AOV, etc. You can keep track of this in a spreadsheet that accumulates over time. Include headings such as Tactic, Date, URL Parameters, Target Page, Content, etc.

• Prior to launching any optimization, setup segments for your top traffic sources. This will allow you to monitor KPI’s across key marketing segments and to create targeted optimizations easily if the opportunity exists.
• Assuming that you have marketing campaigns running at different times of a month, it is important to allow your tests to run for at least two weeks to account for traffic seasonality. Additionally, your site may have different weekday vs. weekend behavior; it is essential that your test overlap at least two weekends.
• Monitor overall and segment performance at 95% confidence (or greater ) paying attention to both conversion rates and revenue per visitor metrics. More on analyzing results in future posts.

Conclusions

Due to the fact that no traffic is created (or delivered) equally, taking traffic inventory prior to launching an optimization is a key step in uncovering opportunities, preventing premature analysis and increasing revenues. Once traffic inventory is taken, segmentation, targeting and granular testing can lead to maximizing each individual traffic source.

Want to discuss in more detail? Drop me an email to discuss.