Marketing is iterative

Great post today from  A good reminder that marketing is never perfect. If you are doing things the same way everytime, chances are you’re not marketing optimally.  I began my career in consumer finance marketing (yes, you can blame those annoying credit card offers on me) and one of the things I loved about that work was how measurable everything was, from individual customer profitability to responses for individual direct mail campaigns.

I constanty hear people bash direct mail as a waste, referencing the fact that they always throw their junk mail away.  But companies like Chase, Capital One, Bank of America and the like have been doing this for so long, believe me – if it didn’t make money they would know. They would know because they can measure response. And among those respondents they can track the individual profitability, and then adjust their marketing accordingly.

Now obviously a lot of banks today are losing their shirts but that’s generally a function of risk management not ineffective marketing.

The other thing I loved about direct marketing in financial services was that our customers were usually in the tens of thousands or more, which meant plenty of ways to slice, dice and test-control, test-control. And with each test you could tweak ever so slightly and evaluate the comparative results. When you get email spam, chances are you’re reading one of at least two or three varying subject lines and/or varying bodiest of text. Likewise with direct mail, you’re probably seeing one of two or three variations in the envelope or some other part of the message. That said, there are some basic rules of thumb to testing your marketing.

1) Don’t change too much. If you change everything at once you won’t know what change impacted what results. Remember this is an iterative process. You can’t learn everything overnight.

2) Similarly don’t target the same customer with multiple offers. I was once in a business where three different marketing manages were targeting the same subset of consumers, and because they didn’t have the appropriate tracking mechanisms, it was impossible to know what tactic drove what action. You also run the risk of diluting your message or scaring customers off with inconsistent/badgering messages.

3) Remember that when testing you should select a sample size that will give you a sizeable response. Rule of thumb is n > or = 30, where n is the number of respondents (not the overall sample size).  So if you expect a 2% response rate, you should have at least 1500 folks in your test sample. Less than 30 data points and your results could be skewed by the ideosyncratic behavior of just a couple respondents.

4) Use common sense. We are bombarded with messages in all shapes and sizes. I see a lot of marketers try to push the envelope with whacky mailers or eye grabbing e-mail subject lines. The problem is that more often than not I know right away those things are spam…so if you’re going to do that, the message better be very compelling. At the end of the day people have become skeptics. Your content should be genuine and the product/solution should speak for itself without all the bling. Your job as a marketer is to clearly communicate that value.

As Peter Drucker once said: “The aim of marketing is to know and understand the customer so well the product or service fits him and sells iself.”

Marketing is iterative if for no other reason than because you are constantly getting to know your customers better.


, , , ,

  1. #1 by Laura on May 13, 2009 - 9:30 am

    Good points and very applicable to market research. One thing I would say is that n should actually be at least 50, and preferably 75-100, for robust results and ability to have a 95% confidence level. Larger isn’t always better (your confidence level only gets incrementally better once you get over 150 and it might not be worth the cost to get those extra people), but if you’re too small, then you wont’ have results you can believe in and compare against.

    • #2 by Kris Kaneta on May 13, 2009 - 9:42 am

      Hi Laura – I agree 50 to 100 respondents would be better. I generally go with 30 because a lot of times marketers just don’t have the sample size to support such a test. If response is 2% and we wanted 100 respondents, thats a sample size of 5000. A lot of times, especially when you move away from mass market consumer goods (i.e. b2b), your entire universe may be 5000 or even less. Say hi to the Deac for me.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: