Following a previous article on marketing metrics, many of you raised the question of measuring marketing impact on the company bottom line.  While the general consultant answer – “it depends” – would hold true here as well, my real answer to most of you in the enterprise software world is simple but disappointing: YOU CANNOT. 
 
I can hear some of you crying foul as you’re reading this answer: “With all the talk about ROI, marketing still refuses to be measured.  What good are all these marketing metrics”, you say, “if you cannot measure the impact on the bottom line?”  Before you get all excited, let me ask you some similar questions: Can you measure the impact of your product development on your bottom line?  Can you measure the impact of every sales call on the bottom line? 
 
Reality is that your company’s bottom line is impacted by too many variables.  Establishing statistical dependencies when many variables are involved requires large samples of data.  In response to a recent IDC survey, only two out of the ninety companies surveyed had overall marketing ROI measurements, and these two were companies with over $10B in sales.  You may have this kind of data if you are a large company or if you sell low-cost software.  But for a young enterprise software company that closes 10-50 deals a year, it’s going to take a while before you have enough data points to establish these dependencies.  In this timeframe, many of the variables are going to continuously change - the economy, your product, the number of references you can provide – to name just a few. 
 
I am also leery when people try to attribute a sale to specific marketing activity.  This may work when you sell online, but in a solution-selling environment, where the sale cycle is 6-18 months long, there are likely to be multiple marketing activities that contribute to each sale. 
 
Does it mean you have to give up measuring your marketing?  Not at all.  Here is what you can do:
 
Define measurable goals
How do you know what these goals should be?  At Cisco, every employee is compensated based on measurement, but the measurement is not the company’s bottom line but rather customer satisfaction rating.  Cisco measures customer satisfaction because its management has a vision that ties it to long-term bottom line results.  Do they have the mathematical formula to prove it?  Maybe by now they do, but probably not when they started.  The same type of vision is what you need to develop for your marketing. 
 
To define meaningful goals, you have to define your own Marketing Impact Model.  This model translates your vision of how marketing will impact the bottom line into measurable metrics.  The model I like to suggest has to do with the level of permission established with your target customers.  For each marketing activity, you can measure one or more of the following:

• Capture: how many new target customer permissions have been established as a result of the activity?
• Maintain: how many target customer contacts have responded to this specific communication?
• Upgrade: how many target customer contacts have responded in a way that gives you permission to take the dialogue to the next level? (we’ll be writing about upgrading permission in one of our upcoming issues).
• Cost: what was the cost of each of the above?

This is just one possible model.  You may have a different one.  What’s important is that you have one explicitly defined, and that you get your entire marketing team and your management to buy into it. 
 
Embark on measurable activities
In the previous article, I argued that every activity is measurable.  It is also true that some activities are more measurable than others.  Online activities are more measurable than offline activities.  For example, e-mail is easier to measure than print mail.
 
With e-mail, you can easily test different messages at very low cost, while printing multiple versions would be costly.  You can measure not only response rates, but also open rates, which tell you how many people read your message and can help you fine-tune future campaigns for better results (see the article Analyze This: Enhancing Email Response below).  And you get reliable measurements within 24 hours, which means you can adjust quickly.
 
What about branding, you may ask, how do you measure branding activities?  To me, permission is the ultimate measure of branding.  Branding is all about trust, and what is a better measurement of trust than permission?  Use your marketing budget to capture permission, and you have measurable branding activities! 
 
Test one variable at a time
Generally speaking, four variables will impact the success and the cost of any campaign:

1. The target audience (e.g., the list you’re using)
2. The vehicle (for example, e-mail vs. print mail)
3. The offer (e.g., white paper, Webinar registration)
4. The creative

To effectively test the impact of any of these variables, you must keep the other three variables constant (unless you have enough data and the knowledge to run statistical regression analysis).  For more on this subject, see the article Test vs. Control.
 
Re-evaluate your marketing impact mix
From time to time, you should go back and revisit your goals and the marketing impact model.  Are these goals still valid? If you’re doing a good job growing your own permission list, you should be shifting more of your activities and your budget from capturing permissions towards maintaining and upgrading them.  As a result, you should also see your overall cost of marketing going down. 
 
Real-life examples
It’s amazing how much you can learn when you follow the process.  Working with a client to promote a webinar, we tried four different subject lines in e-mail invitations sent to a test list.  The results were totally the opposite of what we had expected.  With these results, we were able to generate higher response rates by using the more effective subject line in the invitations sent to the entire house list. 
 
Working with another client, we found out that the cost per response to an e-mail invitation was a hundred times lower than the cost per response to a print version of the same initiation. We were able to figure it out because we followed the rules: we had a measurable goal (registration), we used measurable activities and directed all responses to unique landing pages, and we targeted the same offer and creative to the same list.  This way, we could effectively isolate the delivery mechanism as the only variable tested. 
 
Like any management function, managing your marketing is neither pure science nor pure art.  It requires experience and knowledge to set the vision and the measures to make it a reality, combined with the process and the patience to learn and adjust. 
 
One of our client executives summarized it nicely: “I will approve any marketing request as long as the results are measurable.  I don’t mind spending the money if I know what I’m getting for it and I can use it to improve moving forward.”
 
Let me know what you think.