1. Integrated campaigns
Now that we’ve talked about different advanced analytics methods - let’s talk about analysis options for campaigns that include many different channels, also include “integrated marketing.”
2. What are integrated marketing campaigns?
Integrated marketing basically means: a campaign with multiple marketing channels working together. This could be a campaign driving customer awareness through TV, radio, and paid search advertising. Or it could be a product release campaign including podcast ads, sponsored blog posts, and social media messaging. Marketing Analysts are vital to the success of major campaigns like this.
3. Integrated campaign stages
There are many measurement considerations with integrated marketing campaigns because they tend to be more expensive and complex.
High expense campaigns set targets for each channel during campaign planning before launch. Campaign planning also focuses on predictions, because we can utilize data from previous campaigns to predict which channels and tactics could work best in a future campaign.
During the campaign, Marketing Analysts monitor KPI pacing toward targets.
Post-campaign activities usually focus on analyzing cross-channel trends and assessing ROI impact after the campaign ends.
4. Campaign planning: audience segmentation
Let’s break down some of the most common things marketers want to predict when planning an integrated marketing campaign. A common example could be: “who should we be targeting in a discount campaign?"
As a Marketing Analyst, we can start by asking marketers if this campaign is targeting new or existing customers, specific geographies, an age group, or specific products. By leveraging the skills we learned in the previous lesson, we can segment our historical campaign data around categories based on the answers to those questions.
5. Campaign planning: response models
Once we narrow down possible customer segments, then we can move on to our next question: “who is most likely to purchase when a discount offer exists in a campaign?” To answer this question, we want to use a response model.
Response models classify customers who will likely respond (or purchase) in an upcoming campaign. If we have historical data for groups of people where they did not receive a discount versus those who did receive a discount, we can build a predictive model to understand the change in response probability and recommend the group most likely to respond to a discount offer.
Response models typically use statistical hypothesis test comparisons (validating if results are due to chance) or regression modeling (estimating the strength of the relationship between variables).
6. Campaign planning: choice models
Now that we know which audiences we think this campaign should target and which are most likely to respond to a discount, we can ask: “what type of discount messaging will resonate with customers?” If Marketing Analysts answer this question, they can advise marketers how to structure campaign messaging to further increase likelihood of purchasing. Does it matter more to have the greatest percent discount or special financing or a limited-time offer?
By using a choice model, we can compare multiple offer attributes and rank them by purchase probability.
To make sure we account for discount messaging options, we can use a regression model.
7. Campaign planning: experimentation
Once we make our discount campaign recommendations, we may be asked to plan for the campaign to be executed as a test.
In a perfect world, we could set up a simple A/B test (meaning a randomized experiment) and show different discounts randomly to different audiences; the reality is that running an integrated marketing campaign as an experiment is not always feasible or cost-effective.
We can either recommend a randomized experiment or use causal inference, where statistical models are used to simulate the marketing campaign.
8. Let's practice!
Let’s apply this new knowledge about integrated marketing analytics to a sample campaign plan!