Is It Working? How To Successfully Measure Marketing Campaigns

Ah, the age old question that the finance department is always asking of marketing. How do you measure what you’re doing? How does it affect the bottom line?

Obviously things have moved on since the days of Don Draper and we have some incredible tools at our disposal to measure the effectiveness of marketing campaigns. We now have the capability to see exactly how every minor change, every A/B test, every channel decision, affects the end goals.

But it’s still not that simple. The recent IOS deployment has, at the very least, muddied the digital tracking waters. And existing attribution models, like the ones you’d see in Google Analytics, are flawed ways of measuring and analysing the behaviour of users and therefore the effectiveness of campaigns.

In this blog we’ll go through some of the dos and don’ts of measuring marketing campaigns effectively. Let’s dive right in:

DO: Have clear goals & measurables for your campaign

The first and most obvious thing you might think your marketing campaign should create is return on investment (ROI). However, this isn’t really sufficient as a way to measure campaign effectiveness. For one thing, there are plenty of cases where immediate revenue isn’t the goal of a campaign. It may be app downloads, it may be brand awareness, or it may be likes and follows on social media.

It’s therefore key to set proper goals and key measurables for your campaign. At Growth Division, we like to use objectives and key results (OKRs) to measure the effectiveness of our campaigns.

First, you set a clear, relevant, time-bound, achievable objective. For example: increase the number of items bought on your marketplace business by 30% in 3 months. Under that objective you set out a list of key results that will help you meet it. For example: number of buyer sign-ups, CPA (cost per acquisition), number of items bought per user.

In setting goals, we find it’s always helpful to think in terms of the marketing funnel. Lay out a marketing funnel specific to your business, broadly based on the awareness > consideration > conversion > retention > referral framework. Think carefully about what areas you want to prioritise, and what specific objectives apply.

Once you have your OKRs, it becomes a lot clearer how to measure and understand the effectiveness of your marketing activity. 

DO: Run constant experiments

Our approach to marketing at Growth Division is driven by experimentation. For every activity we undertake, we define it as an experiment with a hypothesis, expected results and time-frame. We then capture the actual results of the activity and close of the experiment, making sure any and all learnings are applied campaign-wide where relevant. We cover this in more detail our articles What Is Growth Marketing and How To Create Your Growth Marketing Framework In 7 Steps.

What it means to run constant experiments is you don’t just start running Facebook Ads and wait for more sales. Your campaigns are in a constant state of evolution, finding new ways to improve your key results and hit your objectives.

By using this experiment-based framework, any campaign you run will by definition have very clear measurables, and its effectiveness will therefore be much easier to measure. You’ll know right at the start of the campaign what needle you’re trying to move, and why you think your marketing activity will move it. 

DO: Make use of all the martech tools at your disposal

Modern marketing activities have a huge array of tools to help you keep track of results and judge campaign effectiveness. From Google Analytics, to the Facebook Pixel, to CRM systems that integrate with all your outreach channels via API.

As an example, if your campaign centres around sending users to your website, who can then create an account and make a purchase, Google Analytics can give you a wealth of information. If you have Enhanced Ecommerce set up you can track the biggest sources of revenue for you in terms of location, demographics and acquisition channel. You can measure the effectiveness of a particular activity by using Urchin Tracking Model (UTM) parameters on the links to your website.

You can even use attribution modelling to get a view of how different channels interact with each other in a user’s journey through the marketing funnel. However – this is flawed, as we’ll go on to discuss in the next section.

The martech tools to use are completely dependent on the sector, product and OKRs you have. If you’re not sure what tools will be relevant or useful to you, talk to us. Building martech stacks for clients is one of our core agency strengths. 

DON’T: Think analytics will give you easy answers

Martech tools are a must-have for any campaign that touches the digital space at all. BUT you can’t always rely on them for everything.

One of the main complexities is accurate attribution in multi-channel campaigns, and this has only got more complicated with the launch of IOS 14, which has been a seismic shift in the way user tracking is done. Pixel-based tracking is, effectively, going the way of the dodo.

When you’re running a multi-channel campaign, and you have an objective such as ‘increase the number of leads coming through my website, you’ll really want to be able to track where those leads have come from.

Most analytics systems, including Google Analytics, tend to focus on ‘last click’ attribution. That is, the system will attribute a conversion to the place a person came from immediately before converting. If they clicked a link on Twitter, and then came to the site and converted, that conversion is attributed to Twitter.

Unfortunately for us marketers, people don’t behave like that. They might read about something in an article, go to the site, get distracted by a dog outside, see a Facebook ad on a different device, click and convert. That’s a difficult journey for analytics to accurately model.

Google Analytics has attempted to help with this by creative different attribution models. Using a pixel, it can track a user coming to a site multiple times and assign percentages of the conversion value to the different routes a user took to the site based on the model you choose. If a user sees a Twitter ad, comes to the site, leaves, then sees a Facebook ad and converts, it might give 50% of the credit to Twitter and 50% to Facebook. But that’s flawed, because who’s to say which ad did most of the work in converting them? And it won’t work if a user uses multiple devices. And it’s even less likely to work if they are an IOS14 user, with pixel tracking turned off. 

DON’T: Make snap decisions

Because of this flaw in attribution tracking, we always recommend not making snap channel decisions based on analytics. If your Google Analytics is telling you most of your conversions are coming through Google Ads, don’t assume you’ll get a lower overall CPA if you turn your Facebook ads off. Your Facebook Ads may be filling the top of the funnel, but Google Ads are getting most of the last click attribution because users are searching for you after seeing an ad.

Now, what you can do is frame this in an experimental fashion. You could, for example, say ‘I hypothesise that Facebook is filling the top of the funnel but users are converting via Google. To test this, I will reduce the Facebook budget by 10% over 1 month. I expect this to reduce overall conversion volume through Google by up to 10%.’ Run the experiment, measure results and apply our learnings.

This is why the experimental approach to marketing is so important. You can’t rely simply on attribution models and pixel tracking to tell you what works and what doesn’t in a campaign – you must test, learn, and validate/invalidate hypotheses.

We hope this has been a helpful explainer for measuring marketing effectiveness. As with all things, there’s no substitute for experience and expertise.

Talk to a Growth Advisor

We create a clear, focused marketing strategy by combining our expertise with your knowledge of your business.