There is no one-size-fits-all approach to marketing analytics. Every company has to decide the best approach for their team based on the complexity of their data and needs.
The key is to choose the marketing analytics approach that most expertly converts your existing data and information into actionable insights.
To help you out, we’re covering four approaches to B2B marketing analytics, explaining what they how, how they work, and best practices.
Descriptive analytics is arguably the easiest approach to marketing analytics. It uses a summary to highlight critical information found in your data. It uses facts and historical information to report on what happened. It does not make predictions about the future—that’s a different approach—nor does it deal in possibilities.
Think of it as breaking down your data into its very essence — key points that you’d want your executive team, stockholders, and other departments to understand.
For example, let’s say you want to give an overview of all the webinars your company has hosted over the last year. You could use descriptive analytics to break the data down into fundamental information such as:
- The number of webinar attendees by month.
- The number of conversions per webinar and per month.
- The top five and worst five performing webinars.
Descriptive analytics is best for situations where you require real-time reporting on KPIs, which means you need to keep it simple. Below are some best practices.
- Collect & Analyze Relevant Data: Develop a shortlist of metrics that are valuable to your company and only analyze those.
- Present Data Concisely and Clearly: Don’t create meaningless charts when a single number would do. And be careful about how you present your charts. Don’t mix up line, bar and pie charts, and make sure scaling is appropriate, so you don’t inadvertently influence perceptions.
- Keep it Simple: Only include the data, charts, and information that are necessary to make your point. Too much information creates junk and confuses the situation.
Where descriptive analytics tells you what happened, diagnostic analytics tells you why. It delves town into your data to find out the root cause of your results, helping you understand the relationships between two or more sets of data. It’s all about uncovering and understanding causal relationships.
For example, let’s say that after your descriptive analysis, you found out that webinar attendance was down 10% from the previous quarter. Using diagnostic analytics, you would try to find out why that’s the case.
- Did you change your ad spend or where you shared information about your webinars?
- Did your website visits or landing page views change during this time as well?
- You would also need to look at data that directly contribute to attendance, including impressions, clicks, sign-ups, views, etc.
Using diagnostic analytics, you might discover that the downturn in attendance was directly connected to the amount you spent on Google Ads. Then, you could create a bar and line graph that demonstrates this connection.
The end goal of diagnostic analytics is to tell a story about why your data is the way it is. To do this effectively, there are a few best practices you’ll need to follow.
- Investigate Carefully: Only investigate data that’s worth a second look and requires analysis. And once you begin your investigation, make sure you look at all the factors—including outside influences—that might have affected the issue.
- Understand Correlation vs. Causation: Just because two sets of data seem to match doesn’t mean they caused each other. For example, your webinar attendance could decrease at the same time as a new product launch, but that doesn’t mean the new product caused your webinars to fail.
- Draw Complete Conclusions: Make sure to state your conclusions clearly and don’t immediately jump to the easiest answer. Check obvious and nonobvious places until you can draw a well-thought-out conclusion.
With predictive analytics, you rely on past data to get more data. It’s not necessarily about predicting the future, but about figuring out data that are not yet available. As long as you have the right data for training your algorithm, predictive analytics can be used to understand:
- Customer interests
- Where a customer is in the sales funnel
- Customer satisfaction
- Lead scores
- And more
For example, let’s say you want to determine if attendees enjoy your webinars and what they think—positively or negatively—about them. Using predictive analytics, you don’t have to gather every comment or customer review and read them individually. Instead, you could use sentiment analysis and a complex algorithm to score each comment positively or negatively based on tone. Then, draw conclusions based on the results of the algorithm.
Predictive analytics is not exact, but it can be used to offer results that could be true based on evidence. The key is inputting the right data and coming up with an appropriate algorithm. Here are a few best practices for success.
- Create a Valid Training Set: You can’t predict unknown data unless you have a solid set of training data to work from. Without enough data, you can’t create a working algorithm that transforms that data into new information.
- Test Again and Again: Predictive analytics relies on optimizing your training data and algorithm. Don’t be afraid to test your results over and over again, improving the outcome each time.
- Recognize When to Use Predictive Analytics: Not all data can benefit from predictive analytics. Recognize when it will work and when the data can’t be used to reach your goal.
The final and most complicated approach to B2B marketing analytics is prescriptive. The goal is to recognize customer behavior and then make insightful changes/adjustments to the marketing process to better meet customer needs. It’s about identifying patterns and then creating rules for actions.
For example, let’s say your CRM monitors your customers’ actions on our website and notices that after reading your blog on product X, customers tend to sign up for your webinar about how to get the most out of product X. Using prescriptive analytics, you might then infer that customers would benefit from self-help resources on product X.
The difference between predictive and prescriptive analytics is that in predictive, you only uncover new data, whereas prescriptive analytics recommends an action. As for testing prescriptive analytics, you’d have to validate your recommendation based on whether or not customers clicked on your self-help resources.
The key to successful prescriptive analytics is measuring the amount of value your business receives based on the recommendations. To make sure this happens, you should follow these best practices:
- Only Make Recommendations You Can Measure: Don’t just identify actions you can take without thinking about how you’ll measure the results. You need some type of feedback mechanism that lets you know that your recommendations were valid and ROI-generating.
- Use a Tool to Help: Prescriptive analytics can be complex; so implement a commercial tool that will help you produce valid recommendations. Adobe, Google, and IBM all offer helpful prescriptive analytics products.
- Try Multiple Strategies: There are many different strategies for finding appropriate recommendations using prescriptive analytics. Try out different tools and approaches until you find one that works for you.
With HubSpot, you can measure the results of all of your marketing activities in one location, generate automatic reports, and use a range of analytics tools.
If you need help getting started with HubSpot, reach out to us at KeyScouts today. Our expert team can help you with your HubSpot onboarding and all of your B2B inbound marketing activities.
Let us help you get up and running with one of the most powerful analytics tools on the market and take your lead generation to the next level.