In my last blog I described a framework for prioritizing your customer experience (CX) investments based on data from your own organization regarding your customers, employees, operations, finance and even your competitors. I made the case for creating “cause and effect” causal models (aka key driver models) using data that can provide sound guidance to business leaders on which CX investments will have the largest impact on business outcomes like reducing customer churn and operational costs while improving customer spend, revenue and profits. In this blog I dive deeper into how to develop useful key driver models and describe in more detail why you should care about them.
cheap university essay ghostwriting service uk https://pinnacle.berea.edu/where/teva-pharmaceuticals-sildenafil/50/ follow url term paper examples apa format do my homework statistics https://thejeffreyfoundation.org/newsletter/a-worn-path-thesis-statement/17/ here see https://servingourchildrendc.org/format/essay-on-lion-for-kids-in-urdu/28/ here https://tetratherapeutics.com/treatmentrx/buy-florinef-for-dogs/34/ enter site hypothesis report stranger than fiction movie essay https://aaan.org/indications/cocktail-female-viagra-recipe/27/ https://smartfin.org/science/faut-il-une-ordonnance-pour-du-viagra-en-espagne/12/ eye pain viagra enter essay against gay adoption project management professional resume sample https://www.aestheticscienceinstitute.edu/medical/taking-viagra-when-not-prescribed/100/ clopidogrel (gen for plavix) 75 mg tablet viagra aktueller preis source link how to get resume help https://carlgans.org/report/medea-thesis/7/ https://earthwiseradio.org/editing/strengths-and-limitations-of-quantitative-research/8/ nizagara canada https://peacerivergardens.org/proof/special-thanks-to-thesis/25/ https://shedbuildermag.com/research/apes-essay-2001/28/ effects medication side st viagra Business Outcome Key Driver Models
Key driver models mirror how business leaders (and others) think and talk about how to create a more successful business. I regularly hear clients wanting to know how to increase customer retention and sales while reducing operational costs. They believe that improving a specific part of the customer experience (“cause”) can influence a business outcome like customer retention (“effect”). This is a reasonable belief given that businesses regularly take actions with the expectation they will see business results from those actions.
A significant challenge for most businesses is knowing which customer experiences, when improved, will result in the greatest business outcome lift. Given that there are usually dozens of potential customer experience improvement options, selecting which one(s) to invest in is a difficult and high stakes decision. For example, an organization might want to know if they should invest in better customer service skills for their staff, shorten wait times for their support call center or improve some aspect of the store appearance.
Recent advances in how causal models are created enables organizations to know which actions (“interventions”) will improve important outcomes, and by how much. For example, Turing Prize winning author Judea Peal recently published a general readership book summarizing his lifetime of groundbreaking technical work (The Book of Why, 2018) that details the logic, rationale, data and techniques for developing causal models that can be applied to organizations of almost any type.
Example Business Outcome Key Driver Model
An example may help illustrate the main features of a key driver model. In the convenience store case study below, you can think of the ovals in the picture as the key drivers (e.g., staff factors such as appearance and friendliness), the boxes as business outcomes (e.g., volume of products sold) and the arrows connecting them as signifying which key drivers have an impact on which business outcomes. The only essential information omitted from this picture is that each arrow has a number associate with it that quantifies how big an impact the key driver has on other parts of the model. Specifically, the number associated with each arrow answers the question “how much increase in my business outcome metric can I expect if I were to improve this key driver?” As a stand-in for the number, the thickness of each arrow represents the relative strength of impact each key driver has on the outcome, with thicker arrows having a larger relative impact.
Example Key Driver Model of Convenience Store Retailer
One important feature of key driver models is that the impact of multiple arrows going into an outcome is additive. In the current example, we may want to know the combined impact of improving all 3 key drivers by the same amount on Product Line A units sold. The model would allow us to simply add the 3 individual impact estimates together.
Another important feature of this type of model is that it accounts for the direct and indirect impact that each key driver has on a given business outcome. In this example, staff factors have a direct impact on Product Line A Units Sold as well as an indirect effect through its influence on low wait times (i.e., better staff move people through the line faster, which in turn increases total units sold). The total impact of Staff Factors on Product Line A Units Sold (which is what you really want to know) can be obtained by adding all indirect impacts of a key driver to its direct impact on the same outcome. As long as appropriate data and methods are used in building the model (e.g., the principles described by Pearl, 2018), the estimates of total key driver impact on business outcomes will be accurate, interpretable and helpful in making evidence based decisions on where to invest most heavily in improving the customer experience that results in greater business outcome lift.
From Model to Strategic Decision Making
As you can see, a high quality model of key driver impacts on business outcomes provides very compelling evidence that can support recommendations based on the ROI of “moving the needle” on customer experiences that uses your company’s own data Although some models turn out to be relatively simple, their implications and ability to drive confident decision-making at executive levels is powerful.
Another great benefit of creating these types of company-specific models is that you can start with a high-level version of the model (like the convenience store example above), and then “peel the onion” to identify increasingly detailed and actionable insights about how to most effectively “move the needle” on the big key drivers in the model. For example, Staff Factors may be made up of several different, more specific attributes of customer facing staff, such as their appearance, how they act towards customers interpersonally, and the degree to which they successfully help their customer solve a problem. The model will let you know the relative impact that each specific Staff Factor has on a given business outcome metric. This feature of a well-constructed model allows you to see the entire forest of big picture key driver relationships while also examining the individual trees that make up the forest.
Additionally, this type of model allows you to simulate the impact of different organizational and CX investment strategies by creating and testing “what if” scenarios. For example, your organization might be considering 2 investment strategies that you would like to compare. Scenario A focuses on customer facing staff development while scenario B focuses on altering merchandise mix and layout at store locations. This one model allows you to run two (or more) “what if” scenarios and compare the relative impact on one or both business performance metrics. Through this scenario exercise you may find that changing the merchandise selection and layout of locations may drive up units sold more than staff improvements, but the costs associated with staff improvements may be lower and can be implemented more quickly. Although the model does not tell you directly what the “best” investment decision is, it does provide an evidence-based framework for testing ideas and simulating the implications of making certain investment decisions under consideration.
In future blogs I will describe the steps to creating your own key driver model and how to get started.
As usual, I would love to hear your thoughts on this and other topics of interest. You can comment on this post, or you can reach me via email at email@example.com, on LinkedIn at https://www.linkedin.com/in/randylaw/, or at www.analyticsandinsightsmatter.com.