What is the value of an email address?

There are any number of ways to acquire an email address.  Change.org or Care2 will run cost-per-acquisition campaigns with you.  You can do online advertising (paid or Google Grant-ed) that drives people to your site.  You can e-append your offline constituents in the hopes of further cultivating your relationship with them.  And there’s organic — getting people to come to your site, then getting to sign on the line that is dotted.

These all have one thing in common: they cost.  They cost in time, treasure or both.  So you need to know whether the effort is worth it.  And for that, you need to be able to put a price tag on a constituent.

This is anathema to some.  Witness our (fake) debate on whether we want more donors or better donors: there are some intangibles that are, well, intangible.

But we are judged against numbers.  Our goal is to raise money and make friends, in that order.  So let’s quantify what we can.

While we are attaching caveats to this, let’s also stipulate that you should do this exercise both for your average email address (since you won’t always know from whence your constituent came) and for as many subsegments as you can reasonable do.  The value of a Care2 advocacy person will be different from an organic advocacy person, which will be different from someone who is looking for information on your site, which will be very very different from an offline donor or a walk donor that you are working to make a multichannel or organization donor.  Each will have its own value and price.

So I’m going to describe the exercise for how you would do a generic email address; the principles are the same for any subsegment.

The first step is to determine the lifetime value of the person’s online donations.  Again, I’m going to eschew attribution modeling as very complex — do it if you can, but if you can’t, you are in the right place.

denslows_three_bears_pg_3

You might think, as I once did, that the way to determine this is to take the online donations you have for a year and divide by the number of email addresses.  However, this ignores that many of your donations are made by people who are not online constituents (and may never be).  So this estimate will be far too high.

You might think, as others I’ve seen do, that you can derive this by totalling the amount given by the amount given to ask emails throughout the year.  However, this ignores that your email may stimulate a desire to give that is fulfilled on another device, another day, and even by another method (more on that later).  Counting just donations given directly to emails will give you an estimate that is too low.

So those are the Papa Bear and the Mama Bear solutions; what does Baby Bear say is just right?  I would argue that you should count the donations given online by those who were signed up for and receiving emails at the time of their online gift.  This too will be an overestimate — you might have received some of those gifts if you didn’t have those folks as constituents.  However, it’s much closer than the Papa Bear model and, as you will see from having run your numbers on revenue per page from yesterday, a constituent gift is far more likely than a person-off-the-street gift.

You also need to factor in the lift that online gives to other channels.  I recently saw an analysis of an e-append that still has double-digit increases in both response rate and average gift of the mail donors four years later.  And this included people who had since unsubscribed.  So properly written and targeted emails can be a strong retention tool.

You can look at your file and see what the offline donation and retention rates are for people for whom you have email addresses and those who don’t.  The challenge is that these are likely to be different types of people.  You ideally want to compare someone to themselves before you had their email address as well as a control audience.

That’s why I like to look at e-appends of the past for this.  You can determine:

  • Value of average donor before e-append who got appended
  • Value of average donor before e-append who didn’t appended
  • Value of average donor after e-append who got appended
  • Value of average donor after e-append who didn’t appended

From that, you should be able to derive the lift that email itself gave.  (If you need the formula, email me at nick@directtodonor.com; it’s a bit boring to go through in detail here.)

Similarly, for events with online registration, the good news is that a lot of walkers fake their email addresses or don’t give you one.  How is that good news?  It gives you a nice experiment.  Take the people who gave you their emails versus those who don’t and their return rates and gifts given/raised amounts.  My guess is that being on your email list should increase both retention and value.  These too can go into the lifetime value hopper.

Now you have a formula to got back to your analysis of pages.  Maybe those advocacy participants of today are likely to be your donors of tomorrow.  Or maybe your Change.org advocates didn’t convert the way you would like in the long-term.  These will help you make choices around investments, pages, and people.  Hope it helps!

What is the value of an email address?

A direct marketing bridge to… cause-related marketing and sponsorship

What, you say?  Corporate is a completely different silo in our organization from direct marketing.  It’s not even like major gift officers where they are working from the donor files we create – corporate relationship folks are working directly with C-level execs from companies, not people who started out as $15 donors.

Au contraire, mon ami.*  Direct marketing can be useful in helping secure relationships with companies that didn’t know they wanted to partner with your organization.

The first step is to append your file with as much data as you can get your hands on, if you haven’t already done so.  You are looking for:

  • Demographic data – age, sex, income variables, etc.
  • Political data – which candidates someone gives to is a matter of public record. This has almost nothing to do with reaching out to corporations but is something you should have on file from a data perspective, as people who donated to political campaigns are significantly more likely to donate (and donate more generously) to nonprofits.
  • Purchasing patterns
  • Interests

Ideally, you would also survey your list(s).  This is done most inexpensively online and can help you get a feel for the demographics of your online supporters and event participants.

From this, you may already see some potential partners emerging.  If your core constituency has an unusually high percentage of people who drive motorcycles, your corporate development folks, armed with these data, can make a more effective pitch to those companies.

This list, and your information about it, is your gold mine for the corporate world.  Assuming that there is not a strictly philanthropic reason for them giving to you, they are generally interested either in what partnering with you will do for their brand among a certain segment or segments of their customers or potential customers or in what your constituents could be persuaded to do with them in a cause-related marketing relationship.

Even in the first instance, your list is your gold mine because companies will assume that if you have, for example, heavy support among 35-55-year-old women, their 35-55-year-old female customer base might think highly about their support of you.

There are a couple of key factors in these relationships, though.  First is never to give up control of your list.  You can allow the partnering company to mail, phone, and even email your list (assuming your privacy policy allows it) with an offer for a cause-related marketing or affinity promotion done jointly with you.  But it needs to be your list, with your control over when and how it is communicated with, with no ability for your partner to simply absorb it into their list of information about their customers or into a prospect list.  In fact, you will want to introduce some dummy constituents into any files you share, even under an NDA and the strictest legal contracts, to make sure a list not used without your knowledge or consent.

The second is to make sure as the nonprofit, you are not responsible for the heavy lifting.  These cause-related marketing programs abound, with people more than happy to give you 10% of their sales, as long as your constituents enter promo code RIVERRUNpastEVEandADAM.  As a nonprofit, you are only ever able to acknowledge the relationship, state the nature of the relationship, and thank the company for their support.  You are not able to, and should not be able to, sell a product effectively.

And any partner that truly values you as a nonprofit will not ask you to do so.  The ideal relationship is one where the company values their relationship with you and promotes it as you thank them and appreciate their support.

Direct marketing can also help you acquire cause-related marketing.  Remember the ability to target specific individuals with your advertising? Look for your corporate sales team’s target list and market your programs and efforts to this audience.  They will think you are massive and omnipresent, when in reality you only could be with their support.

Additionally, your email list can help get you contacts at key companies, by looking at the .com portion of the address.  You don’t necessarily need to have the CEO on the list; the right janitor who believes in your cause may be willing to help you navigate to the right person at their company or help arrange a lunch and learn with the corporate staff.

So your direct marketing list and tactics can help you in the cultivation, success, and execution of corporate programs.  Good luck with this and please share any success stories in the comments!  Thanks!

 

* French for “That’s some straight up bull”

A direct marketing bridge to… cause-related marketing and sponsorship

The basics of direct marketing reporting – part two

Yesterday, we talked about the key metrics you want to look at in Excel – 13-14 indicators that speak to you about progress and testing results.

However, a direct marketing Muggle will look and these data and say “Huh.  Interesting.” This is direct marketing Muggle code for “this is not interesting and it makes me think of my Algebra II class, which was taught by a nun.”

While you will want all of the data, you will want a skinnier, clearer chart for others, preferably with colors that call out what is actually important.  Let’s look at a fairly standard test – your thesis was that extra personalization in the letter would increase average gift versus your control.  Here’s what this could look like:

uglytest

The first thing to notice is that your hypothesis was wrong – average gift didn’t go up.  But now you have another decision – should you pay for the additional personalization in the future?

You, as a direct marketing professional, can read this chart.  The increased personalization caused response rate to increase.  As a result, gross income per piece went up and net income per piece went up.  However, return on investment went down; the additional investment didn’t bring in as much as the investments before it.  What would you recommend to your boss?

This is a judgment call based on your goals for your program.  One good approach would be to call for a retest – possibly with even more personalization or to see if you can get the personalization costs down or different ask strings to try to boost average gift.  This is clearly not in the 95% percentile one way or the other (which are other good fields to add to your spreadsheet when you get more advanced), so more testing would be good.

But I know which one I would mail more quantity of when the next test is done – I would use the personalization version as the control.  For me, net per piece matters more than ROI.  Our donors’ time is a scarce and valuable commodity.  There are only so many times you have the opportunity to get in front of them, so if you have the opportunity to maximize their investment of their time, versus trying to go for cost control in borderline cases like this one.

Charity Navigator would disagree with me, as they focus on cost of fundraising, so that’s another point in my argument’s favor.  Remember the Charity Navigator Constanza test – hear what they have to say and do the opposite and it will be to your benefit.

So now you have your course of action.  Now you have to have other people see it your way.  Time to explain it:

pretty test

The first thing to note is that it’s legible.  The second is quantity and absolute gross, net, and cost numbers are gone.  These don’t have any relevance to the decision over what to roll out with.  If you leave them in, there’s a natural human temptation to think biggest = better, especially when it’s called revenue and has a dollar sign in front of it.  For a layperson, it’s good to eliminate these distractions.

Then we’ve color-coded the winning parts.  Control wins on cost; test wins on response rate and ROI, gross income and net.  This helps draw attention to the salient bits.  It is amazing how much these little steps can help focus minds.

You will note that I left ROI in there, even though it is evidence that does not support the case you are trying to make.  I’ve talked about testing as a central commandment on the direct marketer’s tablets.  But testing is nothing if there isn’t intellectual honesty.  You have to make the case, but also give your team all of the information to challenge you and make your arguments better.

This is usually where the aesthetic marketers get us data-driven marketers.  They tell quality stories based not on what is true, but on what we wish were true.

We must become equally good storytellers, because a good story plus data beats just a good story.  On Thursday, I’ll talk about how to present data in a compelling way, but first, we have to figure out how to measure our metrics.

The basics of direct marketing reporting – part two

Appending data in Excel

One of the primary reasons to have a database is to have data in various tables.  A simple example is a database with a people table – a list of your donors – and a donations table.  This allows for there to be a one to many relationship between the donors and donations.  After all, a good donor will generally donate multiple times.  By using a key in the people table, each donation is marked with that same key – e.g., all donations from donor 12398 have 12398 in their donor field and are associated with that record in the people table.

Excel can’t do this.  As a flat file, all you can represent are one-to-one relationships.  If you had to have multiple gifts in Excel, you would have duplicate donor records for each instance of the gift.

But you can append data to a record in Excel.

A while back, I wanted to do some homemade modeling on a file to add sex in as a variable (male/female in this case, to simplify the model).  Unfortunately, I didn’t have a list of who was male or female in this particular file, so I used US census records to determine how many males and females there were with a given first name.  This led to the odd conclusion that someone who had a name that 60 females and 40 males have in the US is a 60% woman.  Thank goodness I was using this for modeling and not for assigning Mr. or Ms.

By the way, if you were wondering if there were any perfect half-female, half-male names, there were three: Ariel, Hong, and Kris.  Sydney, Robbie, Frankie, and Kerry were almost androgynous, just missing the cut.  The Excel sheet is here if you are curious.

To complete the model, I had to put the percentage of femaleness with each record.  I used a somewhat fiddly function called VLOOKUP.

VLOOKUP tells Excel to look through a bunch of rows to find a particular value in the left column of an array, find it, and return another column in that array.  To wit:

=VLOOKUP(A1, ‘[Male v Female first names.xlsx]Sheet1’!$A:$B,2,FALSE)

This assumes that A1 is the first name you want to look up.  This then goes over to the document called Male v Female first names and looks at the first two columns, A & B.  It searches for the name in column A.  If it finds it, it returns back the value in B next to that value.  So, if A1 is Hong, it finds Hong in the array, sees the value next to that is 50%, and puts it in the cell in question.  The FALSE at the end just means you want an exact match, rather than the closest one.

It is by no means as pretty or efficient as a SQL JOIN statement.  In fact, if you have a lot of these to do, as I did, copy and paste it for the full column, then get yourself some tea and a crossword puzzle while Excel labors like a hamster on a wheel.  But, on the plus side, you don’t have to learn SQL.  You still should, and we’ll cover SQL for direct marketers at some point, but you can’t learn everything immediately.

Appending data in Excel