Using Data in Digital Marketing – Value for Money
The Holy Grail for a Digital Marketer is to know how to demonstrate value for money from Digital Marketing and how to make their campaigns deliver more. I.e. to know which channels and activities are making the greatest contribution to the business outcomes you require, and how to get the greatest value for money from the resources involved.
Marketers faced with this question are likely to start by analysing website traffic to discover how popular particular pages are – and then update content and design to encourage users to stay on the website and obey calls to action. The same may go for social media campaigns with effort going in to create advertisement and post content that maximises audience engagement.
For the Marketer keen to demonstrate value for money from Digital Marketing it is relatively easy to find metrics to help. These might have names such as Page Views, User Sessions, Reach, Engagement, Interactions. A rise in any of these will imply a rise in the number of people aware of and possibly interacting with pages on the website.
But the Marketer reporting such a rise must also be able to say what such a rise is worth to the business i.e. what additional revenue can be expected as a result. Only then can questions about value delivered by time and money spent on Digital Marketing be answered.
Before starting a Campaign
Before committing resources to a Campaign it is worth listing the questions that the Marketer can expect to face – and how they will be answered. Some of the most important are shown below.
1. What return can we expect from our spend on Digital Marketing?
This is the question that any Budget Holder will want to know the answer to.
Digital Marketing is there to influence a desired outcome – e.g. an on-line purchase, an application to study, a subscription to become a new member, an additional recipient for a newsletter.
It is this influence that the Marketer needs to demonstrate. How is this done?
Let’s take a simple example. I have an on-line store selling umbrellas. I want Digital Marketing to help me to increase sales.
Looking back over the past year I can see there was a significant variation in sales – some weeks were great, others less so. I want to know the reasons for this variation to see what I can learn. So I start collecting some data.
Unsurprisingly, I discover that some of the variation in sales can be explained by the weather e.g. when it rained, sales increased. But some can also be explained by the number of paid adverts posted on Facebook as well as how visible my website was whenever people used Google or Bing to search for on-line umbrella sales.
This is using data to help to explain outcomes. You look for the strength of correlation between an outcome such as sales and a range of features relating to all the inputs into the marketing campaign – along with any records from further afield (e.g. rainfall) that may also reflect influence.
Let us say there is a clear correlation between umbrella sales and a particular range of social media ads, along with page views arising from Google Searches plus the use of specific SEO keywords.
The Marketer can then employ those same channels and activities in their marketing campaign confident they should influence sales.
Circumstances might of course change e.g. a competitor selling umbrellas might appear. However, the Budget Holder can be given a Digital Marketing plan that shows what channels and activities will be involved, how much it will cost, and the level of return to be expected. In other words confirmation that they will receive value for money from Digital Marketing.
2. What do we measure to know that the campaign is on track?
We will want to measure progress with each one of the activities that we know has an influence over sales – known from hereon as Predictors.
If we were to see a dip in any of those Predictors it might mean a drop in sales. Also, any unexpected drop in sales might be explained by a drop in one or more of the Predictor metrics.
But there may be additional metrics to use.
E.g. although the number of Facebook Ads was identified as a Predictor, the count of people engaging with those Ads might also serve as a Predictor.
If so, it would provide an additional insight into how well things were going. E.g. we may be on track for the number of Ads being posted, however if there was a dip in Ad Engagement we might be seeing an early warning that something has changed and that the current set of ads are less effective than those posted earlier.
3. How do we improve campaign effectiveness?
Let us say we did see a dip in Ad Engagement and we need to get Ads back on track. We may not know the reason for the dip, and therefore it might not be possible to respond directly to the cause.
There is however plenty the Marketer can do to keep making marginal gains in the effectiveness of a campaign. Planning for these ahead of the campaign will mean being aware of the data that will need to gathered as the campaign progresses.
In the case of our Facebook Ads for umbrellas we can apply statistical analysis to each of the features within an Ad to see how well they correlate with variation in a metric such as Ad Engagement. These features might include key phrases in titles or in the body of the Ad, types of image, positioning of calls to action.
Marketers might typically rely on A/B testing to see whether a change to a feature makes a difference. But this can be a time consuming and expensive process. With analysis of the right data, much can be discovered ahead of this being necessary.
4. Other questions to consider
We know that ongoing use of the channels and activities that serve as Predictors should influence our umbrella sales.
However, it is likely our Predictors will explain only a small part of the variation in sales. If more can be identified it means that we can explain more of the variation in sales, and we will know more about how to influence sales. A further question to ask is therefore:
- What additional features influence sales that we haven’t yet discovered?
If our Budget Holder knows what sales to expect as a result of the spend on Digital Marketing, another pertinent question would be:
- What size budget do I need to achieve a specific sales target?
If we know what spend is associated with a specified sales volume, we can set a benchmark metric which is the Digital Marketing Cost per Sale. This allows our Budget Holder to ask:
- How does the Cost per Sale compare across different periods?
With any use of data, planning is everything
Working with data is often time consuming and expensive, in particular if Digital Marketers have to spend time gathering and exploring a wide range of metrics in order to report progress.
By planning to answer questions such as these ahead of starting a campaign you will have a far greater chance of identifying exactly the data you need with which to answer highly relevant questions about value for money from Digital Marketing, and finding the answers quickly.
Going forward I will be taking each of these questions, plus some additional ones, and looking at the methods involved in finding the answers.
—-
Paul Clarke is the Director of FiguringOutData.com, a Data Analytics Service Provider supporting teams make the best use of data to guide Digital Marketing and Social Media management.
For information about our services and for guidance in the use of data to support Digital Marketing please get in touch using the contact information above.