The Marketing Executives Group International

By: Edward Angstadt http://www.linkedin.com/in/edwardangstadt

Recently our organization reached out to an agency partner, asking them to help develop a new ROI calculator for some of our B2B marketing tactics. Like so many other things in life, this exercise was a reminder that where you end up is heavily influenced by your particular point of view. If I can avoid putting you to sleep with the good, the bad, and the ugly of our ROI experience I’ll consider writing this an effective use of time.
So let’s start off with the agency perspective (or to be fair, that really means our interpretation of the agency perspective).
The good: Our agency did some great things with the information they had available. Other agencies I’ve worked with in the past have also demonstrated a particular knack for putting together a framework that can be used to evaluate performance without having all the details available to a company insider. I suppose that is a function of the nature of agency work.

The bad: The flip side of an agency being able to apply evaluation criteria that can be used consistently across organizations, one-size-fits-all won’t be a tailored fit. OK that’s one of those ‘no kidding’ kind of statements (like Madden saying the team that scores more points tends to win the game). In our case the recommended model was highly sensitive to the number of leads driven by a tactic. Our particular business offers a voluntary benefit solution and our revenue stream is a function of the number of employees working for a client. As any savvy marketer knows not all leads are created equally (either in their probability to convert or in their value – more on that later). For our needs we really need to factor in the employee group size rather than just leads.

The ugly: Here is where we hit some major issues, and again I’ve seen these types of problems arise from other agencies in the past. The return that was modeled was a function of two major components; estimated revenue per lead and estimated cost of similar impressions to evaluate the brand impact. I’ll get into what our internal approach did below but here I have to rant a bit about each of these components. First, the revenue estimate did nothing to estimate the money we make from that revenue stream. Now I can understand that an agency doesn’t have access to a company’s financials to back into the appropriate measure but as a marketer who was in corporate finance back in the day, I am horrified how profitability can be an afterthought. Second, I have seen the cost per impression thing come up time and time again as a proxy for the value to a brand. This is a personal pet peeve; it’s irresponsible to say the value a company will receive is equivalent to what they pay for the media. I am certainly not suggesting that the value of the brand impact should be ignored; I only take issue with the circular logic (think of the errors you would get in Excel).

Now let’s look at our internal evaluation of B2B marketing ROI (again to be fair, it’s a bit tough to be completely objective about our approach). After all, if I didn’t think some of this made sense we would have done it differently or at least not made an effort to tell anyone about it.

The good: We believe our methodology does a reasonable job of estimating the impact to the bottom-line. First we evaluated direct responses hitting our sales pipeline. We use Salesforce.com as our CRM so we estimated the amount of revenue that would be driven by opportunities in the pipeline, weighting those cases against a probability for the current stage of conversion and adjusting for the case size. Second, for the brand impact we used the directly generated responses and applied a multiplier to estimate incremental future business that would be influenced by use of the tactic(s). In our case we used a conservative estimate; effectively stating a level of activity equivalent to only a very small fraction of the direct responses would be realized by future brand awareness. And unlike the agency model we didn’t assume our revenue was all profit (but I’ll admit we could probably still do a bit of work on approximating fixed versus variable overhead impacts).

The bad: Reliability of information on sales pipeline conversion is one area in particular where we could improve our assumptions. Our model is relatively sensitive to the estimated sales conversion rate and the quality of information captured within our CRM. I won’t go so far as to say garbage-in-garbage out, let’s just say anyone who thinks all their data smells like roses is either very lucky or mistaken.

The ugly: We struggle to apply our ROI evaluation model to all of our individual tactics due to overlap in what influences opportunities to hit our pipeline. If we were able to assign multiple lead sources to a prospect and weight the level of influence we believe that would help remove that hurdle. For now, having a structure we can apply to overall activity and to particular tactics knowing overlap exists is good enough as a starting point.

At the end of the day, the methodology we decided to use seems to work well for our needs. We will continue to refine the model overtime, but we believe having a structured evaluation process is significant. Just remember there is always another perspective out there and it might be good, bad, and/or ugly.

Edward Angstadt http://www.linkedin.com/in/edwardangstadt