
Marketing Experimentation is important. When you start a business, or even when you have been running a business a while, it’s often unclear what the best way to reach customers is. The default setting is to decide that because you have some familiarity with a marketing channel from a past life that that is the one you should use.If you do this you leave a lot on the table. Marketing experimentation is important and here is why.
In Traction Gabriel Weinberg identifies 18 different types of traction. Thats 18 totally different ways of getting your marketing and sales messages to customers. Don’t fixate on that number because within each channel there are dozens or hundreds of ways of running a campaign and then thousands of ways of doing the details. That’s a pretty huge number of possibilities to play with. Getting your marketing strategy right depends on pulling all this together
Three Approaches to Marketing Experimentation
There are three ways that you can go about deciding what is going to work best for you.
First you can go with your gut instinct and chose several channels. The more experience that you have, and experience with lots of channels helps, the more likely your gut is to be right. Malcolm Gladwell famously called this the 10,000 hour rule in his best seller Outliers. If you only have 1,000 hours doing something your gut call is a lot less likley to be accurate. This means that it’s usually better to use a different approach that helps validate your gut call.
The next approach is to follow the herd. With digital marketing there is always a hot new trend or a better way to do X. Cynics may call it shiny object syndrome. The approach du jour may work for you but you are relying on the wisdom of crowds to give you a solution when they know nothing about your business.
The third approach is to be more systematic about your marketing. This is what I call marketing experimentation. It’s using some scientific techniques to identify the best approaches as quickly as possible. It is not scientific marketing. (I don’t know what that is but to claim marketing is scientific is to miss out on most human psychology).
Scientific Approach to Marketing Experimentation
The scientific method has three simple steps
- Make a hypothesis
- Do an Experiment
- Review the results
I was taught this in chemistry at school and frankly it was really boring. It was far more interesting to shatter ye more test tubes when the experiment went wrong again. Then again I wasn’t paying for the test tubes and getting it right didn’t make me money. So when I realised this my attitude changed.
Make a Hypothesis
Making a hypothesis is a fancy way of saying
I think that if I do this then that will happen.
Isaac Newton had a hypothesis, backed up with the 400 pages of Principia Mathematica, that if he dropped an apple it would fall to the ground. I have a hypothesis that if I write a blog post every day then lots of people will come to my website and some of them will become customers.
That’s not really good enough. A hypothesis has to be measurable and testable. If I say:
I have a hypothesis that aliens built the pyramids
There has to be a way for me to test that statement otherwise it is an unsupported assertion.
The easiest way to do this is to use SMART. Make your hypothesis:
- Specific
- Measurable
- Actionable
- Repeatable
- Timely
Lets do this for Isaac Newton’s apple.
I have a hypothesis that a 250g apple when dropped from a 1 metre height will always accelerate at 9.8m/s
And now for my marketing efforts in increasing my blog traffic.
I have a hypothesis that by writing an SEO optimised blog post each day and distributing by social media that I will be able to increase daily unique visitors to 1000/day in 3 months.
Can you see that it is really easy to see if these are true or not. In the apple case we need an accelerometer or the other high school physics paraphernalia. In my case we need some analytics software and time.
Marketing experimentation
When we conduct the experiment the important thing is that it is quick, cheap and delivers enough results to determine whether it is effective. When I read the history of science it’s clear that there are two types of great scientist. Great theoreticians and great experimenters. Coming up with great experiments is a real skill.
If you have lots of money it is easy to do experiments. In the Manhattan project they bought an entire desert and set off atomic bombs underneath it to test that they worked. That was expensive. As small business and startup marketers we can’t afford to do that.
What we need to do is to come up with small easy ways to determine whether we can develop the traffic. These tests should be set so that they are within your potential ability. Even if you know nothing about Snapchat you should be able to understand what a new Snap account can do in 3 month and then be able to set goals and targets to achieve that. If you can awesome. If you can’t then maybe that channel should be knocked off your list
I like experiments that can be tested for a few hundred dollars in a few days. The cheaper the experiment and the more dramatic the results the better. If the experiment gives you a great ROI then you can expect that that a larger campaign will give you a good ROI as well. IF a cheap experiment does badly why will an expensive campaign do any better?
Reviewing The Results
The most important stage is reviewing the results. This seems counter intuitive. Surely the most important thing is deciding what to test or running the test itself. No it is not.
The review is the point where you get feedback on your actions and you change them for the next cycle. Say you had a terrible hypothesis and ran a terrible experiment. When you review it you say
- The hypothesis was wrong
- We planned to measure it in the wrong way
- The wrong customer segment was targeted
- We didn’t measure the results because….
- We didn’t count how many times we did the experiment
That requires some self honesty, and a forgiving boss, but when you plan the next experiment your starting point is
- How do we come up with a better hypothesis
- Knowing what we do now what would be a better way to measure the results
- Are we sure that this is the right customer to be targeting? Why?
- etc.
What you are doing is self correcting. If you identify the failures – in execution or in what you are trying to test, the next experiment will be better.
With this toolkit you can then quickly and easily run lots of small experiments to test out new channels. This helps you reduce the amount of money and time that you put into large marketing failures. It also helps you to rapidly test other potential channels at little cost or risk.
Get out here, do small experiments, understand what you learned, repeat, grow quickly.

