Just think of Granny Muffins!

Are you doing a good job? No, I mean a REALLY good job. Doing circulation well means never letting complacency set in, always questioning your assumptions and analysing to the nth degree. BusinessWeek’s Sara Howers poses slightly more than twenty questions that everyone in circulation marketing should be asking themselves.

By Sara Howers

We have what I think of as quite a small circulation marketing team. There’s seven of us and we look after the BusinessWeek circulation (both newsstand and subscriptions) for Europe, Africa and the Middle East. But we are not alone. Oh no. We also have an 8th member of the team, a mythical but often quoted matriarch called Granny Muffins. I couldn’t really explain where she came from, but she fulfils a very special need in these often difficult and tight times. She makes us remember the basics – the rules, the caveats – she makes us question what we are doing, she makes us look at our results from a number of different angles, and she reminds us of our core purpose (after keeping within budgets and hitting our production targets) – namely – TEST, test, test, retest, back test, and test again…… So, often the shout goes up "and what would Granny Muffins do??"

It’s very easy when times are tough, your budgets are tight, and you have to keep giving money back, to forget that testing must still happen. And that it must be statistically valid! Any of you marketers who joined up because you only like the creative side of the job, and don’t like maths, should move on……you give the rest of us a bad name. (And just doing email, because it’s cheap, isn’t the answer either.)

So – hands up now.. come on, I want to see who’s doing a really first class job. How many of you are still testing?

* Do you model your tests before you decide which ones to try?
* Do you lay out all your old tests on the table and see if any of them could be tweaked / combined to make them the winners of tomorrow?
* Do you keep a folder full of ideas always ready for the next campaign?
* Are you fairly certain that your tests will out-pull your control? (& if not, why are you doing them?)
* You do have a control don’t you?
* And you do know what a roll out/balance cell is?
* And you do check all the reports from your merge/purge house to make sure that you have correct representation within your cells, by list and by country etc?
* And do you restrict the % of your list tests to minimise risk within each campaign?
* And have you back tested your last control, particularly if the new control was only a marginal uplift, or if you’re seeing a yo-yo response?
* So, how quickly did your last control degrade?
* Have you checked that your test cell size is still appropriate for your current response rates?
* And you do know how to track response curves (by carrier of choice if necessary) and therefore know very early on whether you’re going to hit your targets?
* And you do have the reports to be able to track all this good stuff?
* And you do have a list history database going back to before the ark? (which is backed up, and in a fire proof safe somewhere)
* And you do use the magic formula to make sure that your test results can be relied upon? (see box)
* Anyone still standing????? Well good for you!

So, how do you test anything at all when you’ve just had your testing budgets slashed? The small, simple, and cheap testing ideas are the best ones to use and do remember that there’s no rocket-science involved at all. The KISS (keep it simple) principle works best! Are you collecting your competitors mailing pieces and also collecting pieces from other mailers? Learn from the best; learn from the ones who still have the testing budgets! There’s really no need to reinvent the wheel and as we all know, there are no new ideas out there.


Are you using as many swaps as you can? Lots of savings and as long as you keep very accurate swap logs, and make sure you have a swap agreement signed (in place of a rental agreement) and seed the lists as usual, you should be safe to do this. Yes, there are a couple of folk out there who have real trouble verifying list balances, but once you know who they are, you’ll just stop doing business with them – just make sure that you can trust your own records 100%.

And when was the last time you tested the frequency with which you can use those lists? Do you have to rest them for as long as you have been? Are they still pulling well? Is the swap or rental partner still mailing their own names? We all get very hung up on recency, but with subscription lists, as long as the age of the record is OK within data protection guidelines, don’t think that you should only be selecting actives, and most recent expires – you’d be amazed at how even old expires can out-pull your other non core and banker lists!

Ah, yes, data protection – are you 100% up to date with the do’s and don’ts – which countries are opt ins and opt outs? Does your merge/purge house have ALL the available Panda files (suppression lists). Does your own database capture and track all the various multiple options nowadays – one simple yes/no flag is no longer enough. If you don’t have a matrix then your suppression file is just "so last year!" And is your suppression list growing – it should be!

And seasonality (well, it’s nearly Christmas)? The old adage about don’t mail in the summer or at Christmas – is it actually true for your audience? When was the last time you checked this? If you talk to the lettershops, they are busy 24/7/12 nowadays – so someone out there is bucking these old but still firmly held onto beliefs.


So you have no money for using a creative agency, so no bells & whistles packs are on the horizon, and no creative prizes going to be won….. what can you do? Have you looked at your own outer envelope recently, considered your indicia, tested local injection, moved the window, used a strapline, considered what the "if undelivered" address could be doing for you? Make friends with an envelope specialist – you’d be amazed what’s out there!

And that business reply envelope – look at other peoples BRE’s (sift through your own daily post at home). You do have a reply envelope in your pack? Maybe you’ve tested reply addressed, versus prepaid reply addressed? Does your BRE have anything on it other than the reply address? A priority message? Some other strapline? Some instructions? Just the hint of a follow up? And does it matter what colour it is? If not, then have you at least ganged up the printing of all your BRE’s for DM, renewals and billing yet? You might save some pennies (as long as you don’t have to pay for storage, your control outer size doesn’t change, and the glue doesn’t go off).

And do you know your P/L for every source, by country and have you looked at your renewals at this level too, to work out your LTV – and have you checked recently with your ad sales team – do they need that level of circ in that country or not? Or do they need some penetration into other, maybe new or emerging markets, so does your list bias needs to change?


When was the last time you did any testing on your renewals? On upselling / increasing term / yield?
* RAB’s (renewal at birth)
* Early renewals (roughly 6 months pre expire)
* Advance renewals (roughly 9 months pre expire)
* Email efforts
* Telemarketing efforts
* Segmentation – even if its at the lowest level of separating out and tracking separately your converters and priors
* MGM’s (member get member)

Many of us have increased our level of outsourcing in recent years, often to help save costs – but if you use these business partners to the max, they themselves will often have very good ideas about testing and pack refinements that you should consider adding into the mix; use them as a partner, not just a supplier who you beat up on prices all the time.

If you have done all of these things, then pat yourself on the back – Granny Muffins is very proud of you – you are a professional and consummate direct marketer….and many of these things are likely to have increased your response/term/yield and/or saved you some money – and if there’s any savings, they alone can form the basis of your ongoing testing budget!

And please don’t ever get despondent if you test in every single campaign and you just can’t beat your control (ideally by 10% points or more). You are doing the testing, you are striving for a winner, you are keeping to the path – you can stare Granny Muffins in the eye and know that she won’t wag a finger back at you.

So, please test, test, test and test again…………………… become a testomaniac!

The magic formula

Sample size/test cell = Z2x R x (100 – R)

Z = relates to the confidence level required (1.281 = 80%, 1.645 = 90%, 1.960 = 95%, 2.575 = 99%)
R = the estimated gross response rate that you think you will get.
E = the required/accepted level of error, expressed as a % of the gross response, being a variation either side of R.

The closer you want/need any "roll-out" of a control panel to be, the lower the E and the higher the Z are needed. If you notice that your roll-out of a new control is significantly lower than predicted, then some part of your calculations is wrong, or market conditions have changed very quickly and very markedly – but it’s nearly always the former. I have never yet seen a roll out perform better than a control panel, but it needs to be close, as close as possible if you want to regularly hit your targets. We all need to be able to know that we can hit or exceed our targets, and if you’re doing your campaign management properly then 9.75 times out of ten, you will.