Estimated Reading Time 11 minutes

In any industry, there are theories, myths and truths. This is also true of Digital marketing.

Sometimes without enough evidence to prove the truths, as humans we would rather pin an unknown myth to that question rather than keep an open mind, or further testing something until a more robust answer comes to fruition.

Our brains are bombarded with more information and advertising than ever. Augment this with research, blog posts and social channels…you get the picture. In short, as marketeers our messaging needs to not only be content relevant but also in time and location.

In 2000, the average attention span was 12 seconds. Now it’s 8 seconds. That’s shorter than the 9-second attention span of the average goldfish. (source)

So how do we go about testing something effectively?

The scientific method is a way of testing and experimenting to try to answer specific questions. This process is all about understanding what you are trying to prove, identifying what you believe will be the answer, testing it and then analysing the results to provide evidence.

• Question
• Research
• Hypothesis
• Test
• Answer
• Analyse
• Experiment

This form of testing is never easy, however it is important to follow the process to call the test scientific. When a test does become challenging, all too often people use anecdotal evidence to prove something is true.

Anecdotal evidence according to the Oxford dictionary “(of an account) not necessarily true or reliable, because based on personal accounts rather than facts or research.”

Therefore if I am basing my Digital Marketing findings on personal accounts, then they may or may not have any truth behind them, the only truth is likely to be coincidental.

It is very hard to tell. Many of your ads may be seen on a daily basis, from traditional channels through to digital interactions. Possibly all have a level of influence, but how can you prove it? The main challenge for marketers is how much influence each ad has on a person’s subconscious decision to buy.

How can we know what level of influence a TV commercial has on a buyer’s behaviour?

To add another layer of complexity. If digital advertising wasn’t running in tandem to the traditional channels, would traditional media convert at all?

Several years back, in 2011 Reggie James MD and founder of Digital Clarity talked about tweeting with the telly on.

People have so many distractions, from mobile phones to tablets and digital personal assistants, how can we expect, or understand if the right messages get through and how do we know what is working and what isn’t?

If we know one keyword had the same volume of clicks last week as this week, the ad position is the same, CPC is the same and we have made no changes to the account or keyword, why is my conversion rate down from 3% to 1.5%?

There ‘could’ be multiple reasons; the weather was different, it was pay day week, school holidays, the news was talking about instability within the country, Brexit, Trump etc..

The problem is sometimes it is hard to know why the change occurred. What if simply less people were buying stuff that week?

So, if assumptions are made, for example:

The weather was sunny, hence more people were outside, and less people were buying my clients products online.
Then surely mobile traffic should increase on sunny weekends (which converts at a lower rate than desktop), as more people are accessing the internet outside. Then if this is the case, the traffic would likely be down on desktop meaning overall conversion rate could be down even though clicks are steady. But this is anecdotal, it’s my account of what is occurring and is not evidential.

Let’s analyse and understand if these assumptions are possibly true.

• We need to understand if sunny weather has previously had this impact on our account
• We need to understand if any locations were down in conversions
• We need to know the weather across different locations at different times of day day of week/month etc..
• We may then be able to analyse to show a trend that every day when the weather is sunny in London for example, the desktop searches reduce and mobile increase
• Once this is established we can match the conversions for these metrics and understand if our hypothesis is correct

Overall – if this is the case, the biggest point is what do we do about it? This is the time to test and experiment further.

A note of caution
With transparency comes scrutiny. If your PPC keywords are not making their target ROI, or there is no growth year on year for keyword a, b or c then the question is why why why? There is nowhere to hide.

Sometimes it feels like online marketing can have “too much” data, that is it hard to know where to start. Google Analytics offer some great insights and here are 3 of the top analytics reports I have previously blogged about.

The bottom line is we, as marketers should be setting our hypothesis, executing, measuring and analysing to make sure we are working where possible with facts, or at least tested assumptions. There are no excuses to use anecdotal evidence when the information is readily available.