The ultimate guide to measuring, analyzing, and improving your marketing with the DMAIC methodology

In this article, we explain the data-driven DMAIC methodology that you can apply for measuring, analyzing, and improving your marketing to help your team work better.

It’s always better to try something once than to read about it ten times. That’s why we’ve provided practical examples of DMAIC implementation in marketing for you to make your first steps using this methodology easier.

DMAIC stands for define, measure, analyze, improve, and control. The DMAIC methodology is applied by managers for constantly improving processes, solving problems, achieving project goals, etc. It includes five stages, and it’s easy to integrate this methodology with other concepts such as lean management.

This step-by-step DMAIC process looks linear, but it’s hard to follow in real life. Depending on the organization implementing it, DMAIC involves tens of specialists, tools, data stores, and small and big processes. None of the stages can be ignored, and everything must be harmonized to work efficiently.

Further, we’ll describe how the DMAIC process works inside the marketing department and why it’s one of the best approaches for improving processes in data-driven corporate companies.

At this stage you have to get a helicopter view of your marketing department and the processes it’s involved in. You can create a processes map with pen and paper or with simple tools like Dia. Your aim at this stage is to:

• find out what you want to measure/analyze/improve, define it as your key problem, find its place among your marketing processes, and find its essential data inputs and outputs of your research subject
• describe the real-life performance of all processes involving the object in question
• motivate yourself by setting up goals for improvement — choose the KPIs that must be reached during the improvement stage
• select the team that will be responsible for carrying out the DMAIC process.

What can an object be in the marketing department? To achieve viable results for your business, you have to choose objects that are directly connected with your company’s business goals.

For example, say your business gets hundreds of new customers and your main pain point is a low level of repeat purchases. You can choose a low level of returning customers as a main problem for the DMAIC improvement process. In this case, you have to describe all objects and processes involved in stimulating repeat purchases — your website, the purchasing process, your customer’s buyer persona, your lead generation advertising, etc.

By describing all involved objects and processes one by one, you won’t miss a thing and can arrive at a full understanding of how your subject of research is connected with other elements of your marketing.

During this stage, you have to collect reliable information on what’s causing the problem. The best way to make this information visible is to build a fishbone diagram.

The fishbone diagram, or Ishikawa diagram, is a cause-and-effect diagram that shows you what’s causing a problem in the form of primary causes (most important) and secondary causes (that lead to the primary causes) grouped by categories.

There’s a special set of categories that may be helpful for companies with product marketing. These categories are called the 8 Ps:

• Product
• Price
• Place
• Promotion
• People
• Processes
• Physical evidence
• Performance

The best way to find out the reasons for a problem is brainstorming with your colleagues. The tools you need for this are Dia (or Google Drawings, if you wish), Google Sheets for minimal statistics, and all the sources you need to prove the causes of your problem.

Continuing with the example from the previous stage, your level of returning customers is extremely low, so you get together your team to discuss possible reasons.

During the discussion, your colleagues tell you that the efficiency of remarketing campaigns is constantly low. You mark this as one of the causes on the fishbone diagram, creating a Promotion category and adding remarketing campaigns to it as a primary cause. In Google Sheets, you add the low efficiency of remarketing campaigns as one of the problem causes and enter the quantity of these campaigns.

You then do the same for each problem you reveal during the brainstorming session and build a fishbone diagram like this:

The next step is to interpret data collected in the previous stage to find the root cause of the problem. The root cause is the biggest cause of your problem. To find and separate your root cause from other causes, you should describe the frequency and importance of the causes of your problem and visualize the results using a Pareto chart.

The most important thing at this stage is to concentrate not only on the quantity of causes but on their level of influence on your problem. You can include importance into your calculations to balance quantity and relevance.

For example, you may see that all 20 of your remarketing campaigns have failed, but you also see that your bounce rate for 18 returning users is somehow higher than for first-time buyers, and you see this as a bigger problem than your remarketing campaigns. So while building the Pareto chart, you assign the bounce rate 100% importance and assign the campaigns 80% importance.

The other way to find out which cause is the root cause is to calculate the cost of or profit lost due to problems. The tools you need at this step are connected with those you’re using for marketing, storing data on events, researching customer behavior, quantity and cost of ad campaigns, etc.

By now you should have complete and unbiased marketing data. OWOX BI will help you build any marketing reports in a variety of dimensions you might need in your department. Find out how the reporting magic happens:

This stage is responsible for eliminating the root cause of your key problem, and all its activities must be totally controlled and planned. Otherwise, everything you did before will be for nothing. Remember that your root cause doesn’t exist in a vacuum, so you have to include related causes in the improvement plan.

One approach you can use at this stage is design of experiments (DOE). It helps you plan solutions even for complicated root causes and can be used for any kind of business.

What practical actions can marketing departments take at this stage?

• Automatically create, update, and upload audiences into advertising services to increase advertising profitability without increasing the budget.

• Implement true ROI calculations to find out the real efficiency of your marketing investments.
• Implement ROPO reporting to find the mutual influence of online and offline ad activities.
• Implement data-driven attribution to find out the efficiency of advertising channels and how ad expenses can be redistributed to gain more profit.

And many others. You should make sure your solution addresses the root cause of your problem. Your solution must result in appreciable improvements of your KPIs values.

At this stage, you have to monitor the situation after implementing the designed solution, make and sustain your improvements. Develop your quality control plan to see what you need to keep and what changes are not working well. If nothing went as you supposed, go back to the Improve stage and look for the reasons for the unexpected result.

Remember that all your results are built on the basis of data collection. Advanced productivity analytics is available only for those who collect raw data and merge it in an orderly fashion with advertising, call tracking, CRM, and other marketing data. If your key problem appears repeatedly, even after improvements, set up data collection with OWOX BI and give your analytics a total makeover.

Find errors quickly, determine their causes, and adjust your processes until the metrics show improvements in the field of your key problem.

Imagine you’ve made an error during the specification phase. If you find it and correct it immediately, the fix will be relatively cheap. If you find the error after implementation, when you’re building reports or even making decisions, the cost of fixing it will be very high.

A human mistake in a report for a small campaign may lead to a poor decision for the next campaign, which may lead to an insignificant budget drain or other expenses. But what if such a mistake happens in reports for 1,000 campaigns with monumental budgets?

One mistake in an analytics report cancels out all your data quality. Also, business and marketing decisions based on such reports will lead to lost revenue, additional costs, or even reputational damage and fines.

Automating reports eliminates the potential for human error and guarantees predictability.

Marketing quality assurance verifies that your data was collected, stored, merged, processed, and visualized without errors or leaks. It also gives you the right to assume that your marketing data reflects the real business situation in your company, meaning your metrics and KPIs are dependable and can be used in planning and decision-making.

To assure the quality of your data, you have to implement constant testing of your data at each step of your data-involved processes. The following image shows how data collection is related to the stages of data testing.

The main stages where quality assurance is necessary are the following:

• Setting up business goals and relevant metrics, at which point you should check and test the logic of your metrics calculations
• Setting up automated unsampled raw data collection to guarantee you’re acquiring complete data from your website and advertising accounts
• Installing UTM tags for structured reporting and future analysis

To automate your quality checks on ad data and free some time for generating insights, leave the auditing routine to OWOX BI Data Monitoring. With this tool, you’ll be able to control the importing of ad cost data and reveal all emerging discrepancies. This will help you keep your data comprehensive and guarantee the quality of your ROI and ROAS metrics.

The success of your quality assurance depends on the tools you use in your marketing. First, you must understand how tools align with your business goals and metrics. For example, if you want to calculate your ROAS based on real dollar values and full raw data from your website, you need more than basic Google Analytics. In this case, you should consider tools like OWOX BI Pipeline.

Read more about raw data here:

How to upload raw data from Google Ads to Google BigQuery

How to Export Raw Data from Google Analytics

Always check twice that you’ve correctly set up your tools. Even the most suitable tool might not fit your requirements if it’s set up incorrectly. Also, be sure that your colleagues who are responsible for data quality have the knowledge and skills needed for conducting checks and testing data quality.

This ultimate guide to measuring, analyzing, and improving your marketing with DMAIC is only your first step toward implementing the lean methodology that works efficiently for thousands of companies around the world. Try the DMAIC methodology today and share the results of your efforts.

Good luck with data-driven problem-solving in your marketing!