Recently, I spoke at the SMX Advanced Seattle conference and had the opportunity to attend the other sessions. One of the sessions I was most excited to attend was titled, ‘What You Should Be Measuring – But Aren’t’.

There were four speakers on the panel, the Director of Marketing at the Rimm-Kauffman Group, the Director of Search Marketing at ZAAZ, the President at KeyRelevance, and the Senior Internet Marketing Manager from Unica.

The panel discussion was good. Below are some of the key points that were made:
- Rimm-Kauffman- you should be managing at the keyword level, but preferably at the ad text level. Ultimately, the program goal is to drive profit. And, there is a difference between brand and non-brand terms.
- ZAAZ- “Models aren’t perfect, they are assumptions based on the most data possible. Models provide direction.” He then went on to describe how they used a model to “score” soft conversions for Ford Motor Company.
- KeyRelevance- you need to track offline conversions because they are “the missing metric”.
- Unica- Search is not alone, you need to be looking at all forms of advertising you’re involved in (online ads, offline ads, relationship marketing, direct marketing, social media) and how they are impacting conversions.

Given that the panelists only had ten minutes to present, they were unable to get beyond the introduction of problems to the point of actually providing tangible solutions. After a Q and A with audience members, no real solutions were given to the issues addressed during their presentations. This was evidenced by two questions left unanswered that centered on the topic of Purchase Path™ (also known as attribution management or engagement management). The questions, and ClearSaleing’s answers to these questions, are:

Question 1: If two ads were involved in a sale, which ad gets the credit?

Both of the ads. Most of the time that is the correct answer – the team of ads that’s responsible for a sale deserves credit, assuming that both those ads are clicked in an acceptable time frame (which I’ll define in question 2). However, there are some exceptions to this rule. If the first ad was a non-branded term and the second a branded term, we would give 100% to the first ad. Our Purchase Path technology has proven that the most paths start with non-branded terms and finish with a branded term. Our research shows that consumers enter the branded term as a keyword simply to navigate back to the site they want to buy the product from (navigational keyword). The decision of where they wanted to buy the product from was made after the first ad, therefore, we give sales, revenue and profit to the ads that did the selling, not to navigational keywords.

Question 2: How much time should be allowed for an “assist” to count?

The ideal way to do this is to use a technology that has a way to track Purchase Path. With that in place, you can then plot your sales on a spreadsheet from the time of first visit to close. We recommend looking at 80/20 rule and figuring out which period of time (first visit to last) is responsible for closing 80% of your deals. This is an analytical approach to answering this question.

All in all, I thought it was a good session and the panelists definitely introduced some new concepts to the audience. I do wish that all four panelists would have agreed with ClearSaleing’s perspective that profit is the absolute best metric for evaluating the performance of online advertising. The overall goal for any marketing initiative is to increase the company’s profits directly or indirectly. If you agree with that statement then you should also agree that the best metric to use with that goal in mind is profit (Profit = Revenue – COGS – Ad Spend). All of the other metrics at an online marketer’s disposal (CTR, sales, revenue, conversion rate, ROAS, CPA) can all appear to be trending in the right direction, but that does not guarantee that profit is heading in the right direction, therefore, the only guarantee to making sure you are working towards your ultimate goal is to use profit as the guiding metric.