Integral Analytics [Introduction]
Concept of Growth and Analytics
The way we think and feel about our existence in the universe largely depends on how our parents and community molded us, combined with our tangible physiological traits. Therefore, what we know and what we believe we should know is always related to the environment in which we grew up. When you think about it, progress in our lives depends on four distinct aspects or spectra of reality across two distinct groups:
- Communal
- Individual
Those four aspects or spectra of reality (Wilber, 2005):
- Individual traits: how you are built, how your brain is wired.
- Individual “intangibles”: attitudes and cultural processes you impose on yourself.
- Infrastructure: the physical, tangible things that enable you to thrive and grow.
- Community and cultural norms: what is just, good, and beautiful.
Figure 1: Four Quadrants — Ken Wilber
Take a career of NBA player as an example: If you are tall, you will have a higher probability of being a basketball player than if you are short. Your physiological traits not only play a large role in your success, but also guide your ability to stay engaged within a given activity. Success breeds confidence, but something else is in play. You psychology, mental aptitude and state of mind play an important role. You will need to be able to stay motivated for long years of practice to achieve life-changing success in basketball or any other discipline. That, however, would be for nothing if your parents could not bring you to practice, buy you a basketball, and provide other necessary equipment.
However, this is a book about analytics…“How does all of this have anything to do with analytics?” …You may ask.
First, let us examine what is needed to deploy analytics successfully in any organization:
- You will need a team or individuals who know how to look at the data to answer questions and how to use tools to make analytics (Individual traits).
- These people will need to be motivated to help the business and do what the business needs them to do, as opposed to do what they want to do (Individual intangibles).
- You will also need to give them technology, software, hardware, and support to maintain it (Infrastructure).
- Finally, you will need the rest of the company to know what to do with analytics, know how analytics can help them address business problems, and how to use these answers to make business decisions (Culture).
None of these four elements can exist without the others. Moreover, if one of these quadrants is out of line with the other quadrants, there will be conflicts that will require resolution before the company will be able to truly transform and advance from one level of analytic competency to the next.
Without excess capacity, there is no growth. This brings us to the topic of analytic competency levels. Many books have been written on applied analytics — everything from detailed books on how analytics can be done in Excel, in Google Analytics, Python, SAS, R and other tools, plus how to think about structuring those analyses “for dummies”, to large-company applications of analytics, and, finally, businesses with thousands of employees and complex business models.
One may ask again, “What is the right way of doing analytics, then?”
- What are the right talents, processes, and technologies to deploy?
- Is running analytics for a large business in Excel better because of ease of adoption?
- Should other tools be implemented to manage versioning better and if so how to streamline the learning?
The answer may actually surprise you.
There is no right way of doing it, and all of those books are mostly spot-on with their recommendations. The challenge that companies face today is creating an analytics program that is right for their business. All aspects of analytics (all quadrants) need to be properly aligned in order to operate analytics effectively (support decision making) and efficiently (keep benefits above costs). This alignment will then create a proper balance to advance analytic organization through the stages of development to reach competency levels that top your sector and industry, ultimately giving the company a true competitive advantage in the marketplace.
The reason all these books are correct is that for the most part they address a specific challenge across the quadrants described above (process, technology, people) at a specific level (or two) of analytic sophistication. The basic construct of this book will be organized around four quadrants and across five stages of analytic organization development.
Those five stages are:
- Experiential
- Numerical
- Empirical
- Statistical
- Integral
These stages have a specific manifestation of quadrants, helping you understand where your organization is today. You will also be able to use this book as a schematic for your company’s specific roadmap that can take your business to higher levels of analytic application.
However, this book is a map, not a terrain. You will need to figure out how to apply the necessary changes in your organization so you can move forward on your journey. This means I will not dive deep into specific techniques, tools, and processes. However, I will use applied examples that demonstrate specific features of the map.
Beyond levels and quadrants, I will also cover important aspects of lines (talents, attributes that I will describe in detail later in this book). These are important channels of growth across levels and within characteristics of each quadrant. Also important are different types of approaches to the analytic organization and deployment. Lastly, there is a critical consideration for evolving thinking about individuals to one about team-based quadrants.
The Scourge of Our Times
Business leaders are surrounded by information that brags about best practices and challenges the status quo in every aspect of business operation, making choosing the best option a little tormenting.
This situation prevails in adtech and martech so the major challenge becomes the identification of the right vendor for the right business need. Industry identifies over 2,000 companies in the adtech/martech ecosystem today, (Lumascape, 2017) a number that is likely an underestimate. Technology has never been easier and cheaper to build. (Business Insider, 2015)
In marketing analytics (where I have the most experience), methodologies are rarely defined authoritatively, but instead arbitrarily pegged against alternative methods. The presence of different technology, methods, processes, and the sheer volume of different options creates a confusing environment where it becomes difficult to reconcile the good players from bad ones — and the right players from the wrong ones in terms of your business fit.
Despite many efforts by great organizations such as DMA, ANA, i-com, MSI, IAB, and others, there is no industry standard in marketing analytics, like there is GAAP in finance.
Another aspect of the challenge in the marketing ecosystem is the human capital management (hiring, training, firing) and larger business-process-change management. Rapid growth and fragmentation of adtech space resulted in a market in which few trustworthy experts exist. Consultants who possess truly cross-functional knowledge to develop a balanced roadmap for marketing analytics decision making are hard to find.
Marketing is especially challenged on the integration and broad adoption of independent data decision systems because traditionally it has been predominantly structured around creative and channel-siloed processes. To do it correctly, you will need to break those barriers in several ways (collaboration, data, and decision making are just a few).
The classic example of a fragmented and confusing vendor tech space is multi-touch attribution industry that is comprised of three types of ventures:
- Specialized companies that build sophisticated models to assign credit to media.
- Agencies, technology providers, and internal teams that may pitch their own attribution solution.
- Fragmentation of competencies: Some companies are well positioned to manage your data and some also have better mathematicians while others are more comprehensive. Still others are able to deploy recommendations in the marketplace, shortening the decision-to-market timeline.
Attribution is a microcosm of a broader martech ecosystem challenge. Organizations must consider answering many questions before establishing a vendor relationship or building an in-house solution.
Important Questions to Ask:
- What solution should you pick for your business?
- How will this decision affect your ability to scale the solution as the business grows?
- What happens if your marketing budget doubles? Would the old solution break?
- What type of talent do you need in order to improve your marketing decision-making?
- How do you reconcile marketing investments against finance and CEO expectations?
- What trajectory of growth can you expect in analytic capabilities?
- Does your current setup cater to your company’s needs?
- What are the next technologies and skills you should think about acquiring?
- Where should you expect the potential problems to arise?
- How do you build a powerful process for marketing decision making and have everyone comply with it?
- How do you build and maintain confidence in your marketing analytics team and technology?
Assertiveness in not buying into the current market fad is the key. Business leaders need to be able to distinguish buzzwords from true market shifts and understand how to apply new technologies to their business. That understanding of value creation via technology — does new tech save money or make money? — should be the main principle for decision makers.
Getting to know that a specific solution can actually work in the organization is a more comprehensive process than figuring out where to put the data or hiring few new specialists. It takes all four quadrants to make a decision about which new methods and technologies to acquire. All four quadrants must work together to ensure balanced and effective analytics approach.
Relativity of Analytic Trade
The question of the right tool for the right job, and the right methodology fit for the problem at hand, is a four-dimensional one. First of those dimensions is the level of marketing sophistication of your company. If you are the CMO of a startup with 3 or 4 marketing channels, there is no need to invest in a sophisticated $200k+ annual attribution project, however you still need to grow and develop CRM capabilities. Every business will find their own exceptions, unique to their industry-specific situation. For example if you are managing a B2B business with few large corporate customers, a good client relationships process will serve you well for a long time before advanced analytics practice is needed.
On the other hand, if you are running a $100 million marketing operation with 50 channels, then there is a need for not only a vendor solution, but also an internal specialist team to organize, manage, and deliver insights for decisions that are complex. Despite what many “AI” vendors will try to tell you, there is no substitute today for skilled data analytics professionals to answer questions about your business. The nuance and specific intricacies of your data and required business context will prevent most automated systems from delivering trustworthy results on their own.
Every company falls somewhere on this spectrum between simple and complex needs for decision support, but two universal truths are important to note:
- It is simpler to set the success trajectory process for analytics than to play “catch up”, especially when you are against the wall, being asked to justify your budget decisions.
- You will need only a fraction of the technologies and skills that the market offers and that the pundits want you to believe you need.
This brings us to the next concept.
Stages of Growth
Every business must go through stages of analytics development to ultimately achieve a fully integral approach to analytic decision making. You can define those stages however you like, and many others have already (Davenport, 2012)or (Kumar, 2017). The realization of stages in the first place is more important than their actual names.
Figure 2: Analytics Stages
Once again, I define five stages as follows:
- Experiential
- Numerical
- Empirical
- Statistical
- Integral
- The Experiential stage relies on understanding data via the lenses of one’s own experience. The business executive has a subjective point of view that infers causality. The analytic process sophistication is limited to the experience of a single observer.
- The Numerical stage focuses on reporting and digitization of business practice. In this stage, operators will use tools to track and report all activities such as financial, marketing, and customer related. The causality is derived based on subjective observation and interpretation of trends.
- The Empirical stage involves granular data analysis, using simple summaries and comparisons to understand trends in business. Causality is derived by comparing correlations and building more sophisticated empirical models that analyze several or even several hundred factors. Factors are then weighed against business goals to improve clarity of decision making.
The transitional stage requires a balance between objective and subjective forces.
- The Statistical stage focuses on the further objectification of data analysis via the use of statistical methods. In this stage, the infrastructure is established to handle complex mathematical and computational tasks. Causality is derived by statistical inference and there is intense focus on improving methods as well as business understanding of these methods and their implications.
Scientific methods support evidence. All budget decisions are based on those principles.
- The Integral stage focuses on full cross-department integration of statistical method and automation.
In this stage:
- Fact determines the decisions.
- Complete automation is sought, and applied, for evidence gathering.
- Analytics becomes integrated within the product, making it not only the core operational advantage but also a part of core product delivery and experience.
- Collegiate-level statistics are now transcended by new applied methods that sometimes can go contrary to a common application.
Unfolding Stages in Your Business
An organization moves on the continuum from the experiential stage to the integral stage as it expands and realizes the needs for more sophisticated analytics, encountering more complex data and business challenges.
At every stage, companies can experience stage-specific roadblocks that prevent them from fully transitioning to the next stage. Sometimes, companies can transition to the next stage in some areas but are unable to move on in another area. This imbalance in growth can create systemic pathology that eventually prevents further progress of analytic capabilities at the organization level. Your company could operate in a more advanced stage in some area of analytics or in a given business line while operating on lower level in another.
Every business and even each team will operate at more than one stage. What this means is that you will need to accept that some parts of the process or the organization are less analytically sophisticated than others. Give them their space and evangelize better approach. For others, perhaps you will seek to learn from their best practices, as you will discover they are at the cutting edge of sophistication. You will strive to operate within the core stage routinely. Remember that any given stage defines an average level of sophistication across the business entity.
Read “Integral Analytics [Introduction]” on a larger screen, or in the Medium app!
