How to scale your marketing with one person

July 9, 2010

Written by: Jeff Zwelling

Recently, we received a credit inquiry from a marketing company that was going to do business with one of our clients. We were familiar with this marketing company and frankly, we considered them to be brokers of junk traffic. They ride along the free games sites, installing their toolbar in the process so that they can redirect users to the target site. There is no doubt that they drive millions of visits, but their traffic is mostly junk: kids and people who don’t even know how they ended up on the target site. For e-commerce sites, this traffic is useless as it doesn’t convert.

Instead of letting our mutual client know that I thought this company wasn’t productive, I emailed them to make sure that they were 1) only doing a test to determine the ROI of this channel and 2) that they had the means of determining the ROI of this marketing partner. Unfortunately, I reached them too late because I learned that they had already entered into a yearlong deal. This surprised me on multiple levels and motivated me to write this column.

From my perspective, here’s how to scale up a marketing program with one person and very little risk.

1) Determine your ROI. This should be close to your gross margin but not always (e.g., if you expect to gain repeat business from a single visitor). Make sure that everyone in your organization is in agreement.

2) Establish that you are going to buy all the marketing that meets your ROI goal and discard those marketing programs that don’t. Programs that are close to your ROI goals should be given a short leash upon which either the vendor or you improve until they reach the ROI goal.

3) Establish a test budget. I’ve often used $5K per week or month depending on the expectations of the category. Most marketing programs can establish what’s going on with a $5K spend.

4) Measure the performance of each program. You will need to use a multi-attribution model or else you will not be able to determine the actual net impact of the test program.

5) Put the programs that meet your ROI goals on auto-pilot. Buy as much inventory as you can. Make sure that it’s clear in your contracts that if the ROI goes down, you can terminate with minimal financial damage.

6) For the programs that don’t meet your ROI goal, either terminate the program or provide the salesperson the data that you’ve collected so that he or she can reduce the price to a point that would make economic sense.

7) Set up email alerts to notify you if the program suddenly goes below the ROI threshold (or suddenly spikes). React to these emails accordingly.

8) Move on to the next test.

Most of my time is spent tweaking those programs that are close to my ROI goals so that they will work. Sometimes it involves reducing the reach of the program to exclude visitors by geography, time, etc.