getty
If conversions are the mother lode of profitable marketing, what’s the best strategic approach to improving them?
I recently asked Razorfish chief performance officer Joey Wilson to share his perspectives on how marketing teams can improve conversion rates.
Paul Talbot: When you dive into a project with the goal of improving conversion rates, what do you look at and what process do you follow?
Joey Wilson: Generally, we start and think in 2 key areas and approaches.
First, it’s always inside out, meaning we start in the middle of the process with search results and product results, which need to be the hardest working pages in show business.
Then we move from there down the funnel to cart and then down the funnel to checkout. We use this approach because there is almost always sub-optimization of flow going on. Our job should first be to do no harm and second to improve results quickly so that consumers find it easier to get their purchase done.
We don’t necessarily think in terms of ‘conversion rate’ because it can be deceiving. Conversion rates could go up, but average order value or yield could go down and/or vice-versa.
We like tracking things like ARPV (Average Revenue Per Visitor) because it’s indifferent to outside forces to a large degree.
Talbot: What’s the best way to measure conversions so that marketing teams get the most precise view of the process possible?
Wilson: We love Average Revenue per visitor because this measure doesn’t care about the amount of traffic coming in, conversion rate, etc. It’s a smoothing guide that we find is much better suited to diving in and generating better results fast.
MORE FOR YOU
Most marketing teams mistake efficiency for effectiveness. What I mean by that is many of the optimization measures are intended to minimize a cost variable like cost per click, or cost to acquire, or a TRP. Those are not measures of business performance or top line growth.
I’d prefer a marketing organization report of cost per $1,000 of revenue generated. Then you can see how both the numerator and denominator are moving at the same time and you can measure your actual yield and then optimize against that. Your test will be incredibly more precise, effective and definitive.
Talbot: Can we look at conversions in the travel category? We’re now seeing additional fees for cancellation insurance and the ubiquitous resort fee. When should these extra charges be presented to the prospect so that conversion is maximized?
Wilson: I personally dislike added fees, both emotionally as a consumer and from a commerce perspective. Setting insurance aside, resort fees are something akin to a hidden fee that should simply be included in the price.
At some level the consumer sees the total cost to buy. Better they see it up front so that we can focus conversion efforts on the most likely buyers instead of wasting time, energy and clicks getting someone far down the path only for them to realize the price is 20% higher than they thought and abandon at the last minute.
This scenario would wreak havoc on any plan optimized against cost per click or cost per lead. You would get a path optimizing in completely the wrong direction, likely.
Talbot: Customer journeys often include an attempt at a small, obligation-free conversion preceding what can be a significant financial transaction. What have we learned over the past few years about designing these funnels and customer journeys?
Wilson: We like trials, promos, and incentives a lot but only when they are used well and differentiated. As the old saying goes, the difference between obsolete and innovative is that obsolete worked at some point.
We honestly think many people have lost sight of the fundamentals of journeys, promotions, trials and ease of entry into the customer base. Sometimes easier is just plain better for both you and the consumer.
Talbot: Any other insights on conversions you’d like to share?
Wilson: Don’t be afraid to actually show that you are generating value, real dollar value. Most things go awry when people do not feel comfortable being held accountable, but without that information it’s impossible to truly get better.
Do test, don’t test everything. We often run into environments where people spend months testing something that could have been done in days and likely had a 70% chance of doing better.
Think about ‘Test Vs. Do’ and how to come up with a framework for that system. I see so many really complicated reporting systems that are impossible in 5 minutes to decipher what really is happening and what to focus on.
