How Heavy is Heavy-up?

We often find ourselves in the situation in which our recommendations require changes that are large in magnitude or challenge the sacred cows of an organization.  In many of these cases there is tolerance for a test- either as a way to double check results before committing to a large change, or to “put your money where your math is” to dispel the organizational mythology.

But then a critical question is “how much should we spend?”  It’s a legitimate question; spend too little and you won’t be able to read it in the results, and spend too much and you’ve blown the budget before the test even starts.

Here’s a quick-and-dirty starting point for spend levels:

A. Start with total sales for the heavy up period (in this case, we’ll assume one month): $40,000,000
B. Insert Assumed Sales Lift (see below for guidelines): 1.5%
C. Multiply A x B to get topline impact of test: $600,000
D. Find appropriate margin rate (gross margin, operating margin, profit margin): 45%
E. Multiply C x D to get margin impact of test: $270,000
F. Plug in assumed Payback of test tactic: $0.35
G. Divide E/F to get required test spend for that tactic: $771,000

A. Use the total sales for the product, channel, and geography that you’re going to measure here.  For instance, to test local radio in Chicago for a laptops available at a single retailer, use just Chicago laptop sales through that retailer.
B. Different approaches have differing levels of precision, based on data availability and analytics, but we generally advise that using anything less than 0.5% is too little, 0.5-1.0% has moderate risk of not being readable, but greater than 1.0% is safe.
C. This is the expected “sales lift.”
D. Use the same margin rate that you will use to calculate ROI- finance can help.
F. If you have prior assumptions for the payback of the tactic you’re testing, use it here.  Otherwise, using lower paybacks is safer.  If you’re not sure, we recommend using $0.35.

As always, don’t hesitate to contact us with questions.  We’d love to hear from you.


  • Jan 29 2011
In item G i believe you mean divide E/F rather than F/G. So my question is why do you spend $771,000 to get $270,000 of gross margin lift?
    • Jan 30 2011
    Yes, good catch, I've updated the post. Regarding the $771k spend for $270k upside question- absolutely! It's a question advertisers and managers wrestle with every day. No subject causes more angst in boardrooms (with or without the CMO present). I'll discuss in more depth soon, but here are some facts: 1) The near-term sales (and profit) lift is only a component of the total upside from marketing. Things like brand awareness, imagery, and blunting competition also happen when you advertise, but aren't in that ROI. 2) Almost no one gets positive ROI. CPG benchmarks put TV ROI at about $0.35 per dollar spent. That industry basically invented marketing analytics, so they get those results and continue to spend. 3) Sometimes ROI doesn't matter- if you're communicating a message (think BP's full page newspaper ads after the gulf spill) that isn't intended to drive sales. 4) Low ROI may not mean "kill it." Sometimes the response to a low ROI should be "fix it." One recent client had a terrible radio ROI, and we knew that sister brands in a similar space had a fairly good response from radio. We advised them to run a bunch of tests to see if they could fix it, rather than simply write it off.