Your marketing campaign went out and you didn’t have any tracking for it. “Crap” you say. Maybe worse, but I am not judging.

One method used frequently to measure the impact of a campaign without tracking is called a double delta.

Although that may sound complicated, what that really is, is a measurement of the performance change before and after your campaign went out compared to a benchmark. Either year-over-year or another product during the same time period.

For example: Let’s say you ran a campaign at your liquor store to sell wine in August. To measure your success you might look at sales in July and compare that to August. If sales in July were \$20k and sales in August were \$25k. You might say your campaign increased sales by 25%. Calculation: (\$25k-\$20k)/\$20k = 25%.

Now, how much of that increase may have been driven by seasonality? Or how much of that was actually driven by other activity? What if July was actually really busy and without the promotions, your sales would have been only \$18k in August?

So you already know the deltas for this year, but one way to account for the seasonality is to look at the same numbers last year. Let’s say last year in July, wine sales were \$18k and in August they were \$21k. The delta last year would have been 16.67%. (\$21k-\$18k)/\$18k = 16.76%.

So with that, we can say, the impact we saw was 25%, but historically the impact we expected was 16.67%. So we made an improvement of 8.33%. And because we sold, \$25k and we can attribute 8.33% of those sales to our campaign, we can say our campaign brought in \$2,083. On a side note though, when making imprecise calculations, I find it wise to round the number. As to indicate that your margin of error is larger than if you had exact tracking.

But let’s say you had a promotion last year too and you know that this comparison doesn’t make sense. One method you might do is use a baseline of a similar product this year. Hard liquor sales during the same time period might make sense. For example, if hard liquor sales were \$11k and in August they were \$10k. The delta would be a 9% decrease for the same time period. (\$10k-\$11k)/\$11k = -9.09%.

Therefore if we assume wine sales would have followed a 9% decrease like hard alcohol, the 25% increase is therefor equivalent to a 34% increase. In pure dollars, if wine sales would have followed the same trend, August would have been \$18.1k, but instead it was \$25k, a \$6.8k increase.

And there you have it, a double delta. Compare the trend of your sales, to the another trend, either year-over-year, or another similar product.

Let me know if you have any questions.