Prescriptive Analytics in Digital Marketing: Optimize Your Processes

A major advantage of digital marketing is the wealth of data resulting from your efforts, something unavailable to marketers until now. Analyzing key metrics (termed KPIs or key performance indicators because they correlate with improvements in your performance) allows you to optimize your ROI and grow your business. Yet, analyzing available data takes a different set of skills that many marketers don’t have. In fact, the image below shows the large gap between the marketing skills commonly available and those firms need for success, according to CMOs (Chief Marketing Officers). This is especially true when it comes to more advanced analytical techniques such as prescriptive analytics.

Prescriptive analytics in digital marketing

Before we start, let’s take a look at the term, which may be unfamiliar to many of my readers. Prescriptive analytics in digital marketing can help marketers make data-driven decisions, reduce inefficiencies, and optimize their strategies for the best possible outcomes. Basically, this type of analysis helps brands optimize their decisions to deliver on their goals since multiple strategies exist to achieve any objective. Choosing the right one offers the best chances for building strategic competitive advantage but represents a difficult analysis challenge. It’s a valuable tool in the fast-paced and data-driven world of digital marketing, allowing for agility and adaptability in response to changing market conditions.

Despite the value, we see that few businesses use more advanced analytical techniques, favoring simpler (and less challenging) analysis techniques, such as trend and correlation analysis (see image below). Likely this situation results because firms can’t find analysts capable of conducting more sophisticated analysis and/ or don’t appreciate the return offered by using more advanced analytical techniques enough to pay the higher wages commanded by trained business analysts.

Image courtesy of Forbes

How analysis aids decisions for higher returns

First, let’s take a look at why metrics matter in running your digital marketing campaigns:

1. Organizations using analytics to monitor their business achieve 5-8 times better ROI than those who don’t.
2. In one survey, 64% of marketing professionals agreed that marketing analytics are crucial for success.
3. Prescriptive analytics can raise same-store retail sales by 2-5%.
4. Using advanced analytics, including predictive and prescriptive analytics takes skills businesses need. 60% of businesses need MORE skills in business analytics.

You’ve heard me talk about the difference between descriptive, predictive, and prescriptive data on this site before.

Descriptive analytics, such as #visitors, #fans, and even virality estimates are useful but are limited in their ability to help you make decisions. By definition, descriptive analytics are a historical record of what happened and assume causality between actions and results that are often inaccurate. Predictive analytics go a little farther and predict outcomes. For instance, you can predict your # of sales based on visits to your site if you have prior data to build the model and understand the variables that generate sales. Of course, even armed with a model, you must still make accurate assumptions of future values for these variables, which isn’t always easy. Prescriptive analysis goes even further to help you optimize your returns by finding the BEST solution to a business problem.

Today, we’ll discuss prescriptive analytics used to optimize your business processes — and your profits.

Prescriptive Analytics

Many businesses face problems that require them to optimize their processes given the constraints faced. Prescriptive analytics help you allocate scarce resources, optimize market outcomes or limit risks. Some examples include:

1. With that all firms face a finite amount of resources (time, money, machines, human resources, etc), how much of each product should you manufacture (purchase) to optimize your profit, given each product incurs set costs and returns a set profit? Can you sell the estimated optimal number of products?
2. Should you make or buy products and when should you make/ buy in order to optimize your profits?
3. If you have prospective buyers who reflect certain patterns of buying, how can you optimize your offering to each segment?
4. How should brand managers balance promotions, advertising, pricing, and product assortment to optimize revenue in their retail operations this week?
5. Businesses face a large number of channels for products and promotions, yet they can only support a small number of these options. Which combination of channels will generate the highest ROI?

The outcome of optimizing your processes is it determines the best possible plans, provides alternatives and tradeoffs for various solutions, and can respond to changes in your strategy and/or environment to determine the effect of these changes on your returns.

In digital marketing, you can use these optimization tools to determine how much to spend on advertising (or incur other costs) on each platform, for instance.

The specific optimization tool we learned was CPLEX from IBM. But, there are other solutions for optimizing decision-making (see the section below that contains a list compiled by Technology Advice.

What optimization does

Optimization uses data, objectives you want to optimize, and constraints on resources to determine how the resources should be allocated. Let’s take a simple problem.

Let’s say you set a goal to respond to all comments posted on your social media platforms within 6 hours.

You need to collect data reflecting the average number of comments achieved on each social media platform. You can even arrange this by hours of the day and days of the week. Estimate the amount of time necessary to respond to each comment. Now, define your resources — the number of people you have available and their work hours. You can even define which employees are able to complete which tasks — ie. maybe some of your customer relations staff can only respond to comments on brand 1 while others can respond to comments on brands 1-3 (of course, you’ll also need data on the relative number of comments related to each brand). Or, maybe you have employees who only address customer service issues that are relatively simple while others deal with other types of comments or more complex questions.

Using prescriptive analytics, you can now determine how to deploy your staff to meet your goal of responding to comments within 6 hours. If you don’t have enough or have too many staff dedicated to this task you have the information necessary to adjust staffing to meet your goal.

Let’s assume you use a chatbot (an automated AI tool to answer queries) posted on your social media platforms. While chatbots can reduce costs and increase the availability of your company to address queries, they’re only trained to answer certain types of relatively simplistic queries. You can use prescriptive analysis to determine which tasks are assigned to your chatbot and which to human resources. You can further use prescriptive analytics to determine which queries represent an opportunity for further training your chatbot because the cost of training is less than the value your firm receives from reduced reliance or better performance over human assets.

You can do the same kind of analysis with content creation to optimize returns. You can optimize the number and/or type of posts that optimize your market returns based on issues such as the costs associated with creating content and returns from creating content such as increased sales, improved SEO, etc. You can similarly calculate where to dedicate your content marketing efforts by optimizing the number and type of posts contributed to each social media platform you use.

Below are some other aspects of social media and digital marketing you might want to optimize using prescriptive data analysis.

The process of prescriptive analysis

Of course, conducting prescriptive analysis involves a process with a beginning and an end. Don’t try to skip any of these steps in developing your program and do each step thoroughly before moving on.

1. Data Collection and Integration: The foundation of prescriptive analytics is high-quality and well-integrated data. Digital marketers need to collect data from various sources, such as website analytics, social media, email campaigns, and customer databases, and integrate it into a single source for analysis.
2. Predictive Analytics: Before providing prescriptive recommendations, it’s often helpful to use predictive analytics to forecast future outcomes based on historical data and trends. This can help in understanding what might happen if no changes are made.
3. Objective Setting: Clearly define the marketing objectives and key performance indicators (KPIs) you want to achieve. This could be anything from increasing website traffic, improving conversion rates, or maximizing return on ad spend (ROAS).
4. Advanced Analytics Tools: To perform prescriptive analytics, digital marketers use advanced analytics tools and platforms. Machine learning algorithms, artificial intelligence, and statistical models are often employed to make sense of data and identify optimal strategies.
5. Segmentation: Segmentation is key in digital marketing, and it plays a crucial role in prescriptive analytics. You’ll want to segment your audience based on various characteristics, behaviors, and preferences, as the optimal strategy may differ for each group.
6. Scenario Analysis: Prescriptive analytics often involves creating various scenarios to understand the impact of different strategies. This might include A/B testing, marketing mix modeling, and simulation to see which actions are most likely to yield the desired outcomes.
7. Recommendations: Prescriptive analytics provides actionable recommendations based on the analysis. These recommendations could include adjustments to advertising budgets, content strategies, targeting parameters, or even the timing of marketing campaigns.
8. Optimization: Continuous monitoring and optimization are critical. The marketing landscape is dynamic, and strategies that work today may not work tomorrow. Regularly reassess the data and adjust your marketing strategies as needed to align with your objectives.
9. Testing and Learning: A/B testing and experimentation play a significant role in prescriptive analytics. These allow you to test different strategies and determine what works best in a controlled environment.
10. Measurement and Feedback: After implementing the recommendations, monitor the results and gather feedback. This data can then be fed back into the analytics system to refine future recommendations.

Tools for prescriptive analysis

A variety of software tools exist to help organizations optimize their performance with prescriptive analysis. For the most part, all these tools are fairly expensive and none are plug-and-play, which means you still need experienced staff with the appropriate analytics training to analyze and interpret the results from the software as well as managers who can take the insights and translate them into actions that improve performance.

Here are a few of the options for doing this type of analysis, according to Technology Advice.

• IBM Decision Optimization — Best machine learning
• Alteryx — Best end-user experience
• KNIME — Best data science flexibility on a budget
• Looker — Best for data modeling
• Tableau — Best for data visualization
• Azure Machine Learning — Best data privacy
• RapidMiner Studio — Best data mining and aggregation

I’ve used a few of these, specifically the IBM and Tableau tools. I love the visualizations available through Tableau and, in my experience, these visualizations are critical for non-technical managers to view the benefits of implementing the solutions derived from the software.

What do you think?

Of course, this is a VERY different way of approaching digital marketing analytics. It may not be everyone’s cup of tea, but I hope you see how prescriptive analytics helps run your business and optimize your outcomes. If nothing else, I hope you’ve gained an appreciation of metrics beyond simple descriptive analytics in making the most of your social media marketing.

For a small business, the expense of hiring staff trained in the latest business intelligence tools and techniques may prove too much, especially when combined with the cost of the software itself. These businesses may find it more palatable to hire a third-party analytics firm to conduct periodic analyses of data to develop prescriptions for the upcoming period. Unlike other types of analysis, namely descriptive, prescriptive (and to a certain extent predictive) analysis don’t require an ongoing effort to analyze data. Hence, hiring a third party once a year or so can provide insights at a reasonable cost.

Need marketing help to support business growth?

We welcome the opportunity to show you how we can make your marketing SIZZLE with our data-driven, results-oriented marketing strategies.  Sign up for our FREE newsletter, get our FREE guide to creating an awesome website, or contact us for more information on hiring us.

Hausman and Associates, the publisher of MKT Maven, is a full-service marketing agency operating at the intersection of marketing and digital media. Check out our full range of services.