Ask an Expert: Measuring Email's True Sales Impact

5/03/11

This week’s Ask-an-Expert question is: It's been fairly well established that email drives sales in other channels, ie. email goes out, inbound-call-center sales rise. How do you quantify the effect email has on other sales channels?

Once again, we turn to the experts on Bill McCloskey’s Only Influencers discussion list, a private forum for digital marketers:

Rory Carlyle, director of marketing, BombBomb:

I'll speak to inbound calls because I ran into this dilemma a short time ago. It's actually super-simple.

Market a specific phone number for the email channel. Most times an 800 number is easy to acquire and get tracking on. Forward that number to the call center and put a new referral source in CRM for contact records. Every time a rep answers the phone, they ask a quick, "How'd you hear about us?" and – BOOM: Traceable call volume and conversion, based on email marketing.

If there's desire to get even more jiggy-with-it, there's opportunity to assign a different number and referral source to each sub-channel within email marketing. So; sales, newsletters, retention, giveaways, etc. all get measured for ROI via phone logs and CRM attributes. Plus, the typical statistics contained within the ESP and website analytics tools can be correlated.

See? Simple. :-)

Jordie van Rijn, email marketing consultant, eMailMonday:

All channels have an impact on each other. In a cross-channel world your potential buyers or current customers hop from one channel to the other and if all goes well get a great overall and consistent experience from those interactions.

Most customers will have contact with a company at least three to five times before buying. This might be why email marketers are now also web- mobile- social- instore- ebusiness- pr- search- and "whatever you can think off" marketers, even if their specialty is email. But how to measure the impact email has on these channels?

1. Measure directly

Some of the traffic can be directly measured. Think about "call-me-now buttons" for telesales and coupon codes for traffic to the physical brick-walled shop. You can also "label" the incoming traffic, by for instance using a different (personal) URL if its web-based or a different phone number or special shopping hours/exclusive events for the physical world. That is the most reliable/direct way for promotions but it doesn't measure for instance extra service calls you get.

2. Spikes in stats

Next to that you can check statistics and if there is a spike (or lower activity) as opposed to normal activity, attribute that to "extraordinary marketing efforts", including all outgoing communication like email newsletters.

That, however, will only work if there is a control period or base level established. You cannot compare when you don’t know what is normal. It might be easier than you think though as specific promotions can more easily be put in a different bucket and measured separately. This will only give you global statistics though, as you cannot rule out outside influences like competitor promotions and the weather.

3. Asking.

You could ask the simple question, "where did you hear from us (last)"? Again not bulletproof.

The answers might not all be reliable, people still but email on the pile with "Internet" or "website". Buyers usually have multiple contacts and heard from you via multiple channels before they decide to buy, but it does make it somewhat more insightful or give you an indication to go and measure more accurately.

If you are only measuring the impact of email, you might want to ask "are you a member of our email list?" and grab some new opt-ins along the way if they aren't.

4. Control group:

Use a control group. Send out a promotion, but don't send it out to a randomly selected part of your email list. This way you can measure the impact, also over multiple channels. You will have to be able to identify the clients at the other channels though, if you use a client number or ZIP code that could be an identifier.

You can't go cheap on this and say your non email list members are your control group. Because probably the people on your email list are better (or at least a different type of) customers, so then you would be comparing one type of apples with a different type of apples. ;)

5....and then there is always the option to not measure at all. A lot of companies do it that way, partially because they don't know how, partially because it would be too much of a fuss to get it technically arranged (to identify the customers) and partially because the added value of knowing and going through it with a toothcomb is too small to make the effort.

Happy cross channel measuring!

Bill Kaplan, CEO, FreshAddress:

Revenue attribution among one’s various marketing channels is easier said than done for most companies. The accurate way to do this is as follows: perform a regression analysis using your investments in your marketing channels as variables in the equation and having your sales/donations on the other side of the equation. By holding all of the variables constant except for the one you’re interested in measuring, you can determine the effect that individual variable (e.g. email marketing) has on your sales/donations.
 
For those companies unable to perform this type of analysis, a simpler method is to include a trackable coupon (usable across all of your sales channels) in one of your email programs. 

Tracking results from the coupon redemptions will give you a pretty good idea of the lift across your channels.  Of course, this method is limited by the fact that your results in this instance are only measuring the lift obtained as a result of email marketing in conjunction with a discount/offer coupon, which won’t tell you how your email marketing performs in general (i.e. without a coupon enticement).
 
Rougher measures are simply to track web site traffic as it relates to email sends.  Studies that have been done by Bill McCloskey’s Email Data Source and others clearly show the high correlation between email programs and the number of web site visitors in the following hours and days.

Luke Glasner, principal, Red Pill Email:

To answer this question we should design a test to allow us to measure the affect of email on the other medium. To begin we would draw out the basis of our 2 x 2 factorial test design to set up the test. Since I know a bit about both search and email, I will use those two. You could do all three, but it creates a lot more complexity.

Control Unit – No Email No Search

This is what happens if we do nothing.

Panel 1 – No Email, Search runs

This is the affect of search on sales

Panel 2 – Email runs, No Search

This is the affect of email on sales

Panel 3 – Email & Search run together

This is the affect of running two at same time.

 I would then take results and do the standard interaction calculations. That would give me the relative effects that I could then use to do weights for sales attribution when multiple channels are involved. If you have more than two media running, then you have more complex designs, such as a 3 x 3 factorial design, which is a lot more complex.

Justin Premick, director of education marketing, AWeber Communications:

In principle, I agree with Rory, although I think this approach has its limits.

For businesses employing a limited number of channels, where it is reasonable to think that customers will be exposed to your messaging on a limited number of channels (ideally one, if you want your measurements to be as accurate as possible) in a given period of time, using a unique phone number for a call center, or a unique URL or coupon code for email subscribers, can work.

But consider a brand employing numerous channels: email, catalog, retail stores, television, radio, online display ads, offline display ads and social media.

Over a long enough time, it's reasonable to expect a potential customer to be exposed to a number of these channels (perhaps all of them) prior to making a given purchase. How confidently, then, can we say that email added $X value or X% conversions, based on what number someone calls into? (Or even what coupon code they use when purchasing?) How much of that value should be attributed to the other channels?

This is not to say that we cannot measure the email channel's contribution to sales. But for a lot of brands, a unique phone number, coupon code or Google Analytics parameter won't tell the whole story.

So how do we capture the overall effect of email on our bottom line?

One way is to use holdout groups to test the effect (as suggested here), but even that may not give a true accounting, depending on how long after sending you believe an email can influence sales.

There's a trade-off to be made here between accurate statistics, and quickly available and actionable statistics. The preferences of the marketer and other stakeholders will determine whether holdout groups are practical (and if so, how long those holdout periods should be allowed to last), or if more direct-sales measurement tools such as unique phone numbers are the way to go.

Dave Hendricks, chief operating officer, LiveIntent:

There is a great service called Convertro that can quantify 'view-through' and other email related conversions.

Email often starts the discovery process and dubious clicks from search or blogs etc. get the credit.

Convertro helps demonstrate the value that email provides in the equation. For example, some transactions take more than a single click (they are not impulse buys).  It may take someone 30 days, some site visits, a call or two before a conversion happens. Convertro helps ensure that email gets 'some' credit for the initial discovery, and can manage how much credit email gets once the sale has been executed.

Comments

Show: Newest | Oldest

Post a Comment

Your Name:
Subject:
Comments:
Verification:	Please type the letters in the image above

Terms: Feel free to be as big a jerk as you want, but don't attack anyone other than me personally. And don't criticize people or companies other than me anonymously. Got something crappy to say? Say it under your real name. Anonymous potshots and personal attacks aimed at me, however, are fine.

Posted by: LaTonya Johnson
Date: 2011-06-07 17:43:06
Subject: Excellent Info

This is exactly what I was looking for. Thanks!