The poet John Godfrey Saxe wrote a poem on six blind men who went to observe an elephant. Each blind man stumbled onto the elephant and whatever part they grabbed on to became their perception of the complete animal. So the man who grasped the tusk of the elephant claimed the animal was like a spear, and the man who held on to the tail claimed it was like a rope. They were all correct… and all off the mark. I was reminded of this in a recent conversation. Most digital marketing managers rely on ROAS and CPA to assess online marketing performance. These are the conventional marketing metrics used by old school tools like Omniture and Google Analytics. OK, but not adequate to multi-channel, multi-touch digital marketing. In fact it can be outright dangerous to your bottom line. Read on.
What is Return on ad spend (ROAS)
Definition: Revenue divided by spend. For example, if your revenue is $50000 and your advertising spend is $10,000, ROAS = 500% (50,000/10,000 x 100).
What is Cost per acquisition (CPA)
Definition: The total spend divided by conversions that are attributable to the ad. A conversion, or acquisition, is the goal you want your customers to hit after they click your ad. For example, it could be purchasing a product or submitting their names through a “contact us” form. If your advertising spend is $1000 and you gained 100 purchase transactions through the spend, the CPA = $1000/100= $10/sale.
The problems with ROAS and CPA
Two reasons why there are problems with focusing on ROAS and CPA to measure marketing effectiveness:
- “If all you got is a hammer everything looks like a nail”: If performance is measured on CPA, marketing behavior veers to discounts and promos. Be it email marketing, paid search or display, the promotions veer towards discounts. It’s great to have the equivalent of doorcrashers but as the example on the right shows, all retailers are doing the same. I attribute it to the marketing manager’s metrics being tied to the ROAS and CPA. So there’s the slew of discount ads and promos – with the goal being to convert as many as possible for that campaign. A shopper is more than a big dollar sign (well, that too) – but any good salesperson knows that badgering and hustling only go so far.
- They focus on sales, not marketing effectiveness : Business school 101 teaches the difference between selling and marketing. If anything, marketing is about reaching the customers and building long lasting relationship. So if the metrics ignore the journey and focus on the end result and the budget, the measures are skewed towards selling performance.
Consider the marketing environment illustrated to the right.
- The schematic illustrates the movement of the shoppers through three stages – Awareness, Conversation, Purchase Triggers – on the path to a Sale.
- Each circle illustrates a touchpoint in the shopping journey.
- The two circles “G$” and “B$” indicate two distinct channels used at a particular stage of the shopping journeys.
- Of the four shoppers illustrated, two shoppers proceeded all the way to the sale. Each sale was for $10.
- Of the four shopping journeys, the two that did not convert stopped at the Awareness stage and never proceeded to the Conversation phase.
Since there are two channels in use “G$” and “B$”, the marketing manager could choose to review the respective ROAS.
- G$: ROAS = 1/4 = 0.25
- B$: ROAS = 1/4 = 0.25
Effectively, both channels are equal value(!) to the marketing manager because their respective ROAS is identical. The story is no different if using CPA as the metric.
The Problem: Neither metric registers that the channel “B$” successfully moved the customer to the next stage in the funnel for both shopping journeys.
The Solution: Recognize that the savvy shopper has a multi-touch digital path to the sale and measure accordingly. Here’s how to go about it.
Measuring channel effectiveness in Multi-touch shopping journeys
- Break out shopping journey across funnel stages. We recommend the following
- Purchase Trigger,
- Repeat Purchases
- Establish a channel effectiveness measure A(.) that estimates the likelihood for that channel to move shoppers to the next stage in the journey. A simple probabilistic measure in the example above shows that the channel effectiveness for the two channels is as below.
- G$ measure = 2.5
- B$ measure = 3.5
- Use the measure designed above to isolate channels and campaigns that are effective at specific stages of the shopping journey.
If the above calculations seem too onerous consider that the alternative is to expect the equivalent of a golfing hole-in-one with every discount campaign. The marketing manager I referenced in the opening was frustrated why his campaigns this Thanksgiving were not as effective as previous years.
Is it any surprise that with every retailer doing the same and all advertising verbiage using versions of “Free” and “Cheap”, the shopper has tuned out.
Watch a demo of the multi-touch shopping attribution app
Not sure how to set up your multi-touch funnel analytics. Contact us through the link below to see a no-obligation demo on Polytab revenue attribution app.
Image of Blindmen #6 by Mike Kline used by permission under Creative Commons license.