This blog commences with the SEVEN types of Go2Cab Advanced Analytics. The five actions essential for successful transformation follow the list of Go2Cab offerings in advanced analytics. The context of this blog is at a point in time beyond the construction of the business Target Operating Model, where several critical elements are well-defined. Such components include the identification of your ultimate Customer, the Value Chain, required Services, and Business Capabilities needed to deliver the Services where the outcome of one or more Services will “Add Value,” which, by design, will address one or more Customer needs.
Go2Cab shares its hands-on experience to highlight simple actions you may wish to consider when dealing with Analytics, whether in general or specific to digital reform and transformation.
Business Analytics Go2Cab Offerings:
Go2Cab can help you make sense of your data to ensure the outcome is actionable intelligently. We offer analytics in SEVEN significant areas applicable to various industries such as cloud, transport, portfolio and project management, software engineering, digital transformation, mining, food, safety, health, education, and defence, to name a few. The seven major areas include:
- Numerical Analytics involves using numbers to represent data such as the time it takes to complete a task, your daily weight loss, the tonnes of freight or material hauled from origin to destination, the time it takes to clear an incident or close a ticket in a call centre environment, journey travel time, and so on.
- Categorical Analytics is when your data is judgmental or opinion-based, like data collected via surveys where you answer questions with a selection from “Excellent” to “Very Poor” or from “Very High” to “Very Low”. Another form of answer is a selection of a colour. For example, “Red” may represent very “High Risk” or “Very Dangerous”, and “Green” may represent “Very Low Risk” or “Very Safe”, whereas other colours like “Yellow” and “Orange” are in between the “Red” and “Green”. Categorical data is frequently used in surveys, whether in food, health or any service provided by an organisation where the organisation seeks feedback from the Customer.
- Text Analytics involves data representing textual reports, feedback, reviews of reports, or responses to project proposals. Another form of text analytics applies to feedback or comments made by people via social media like Twitter, LinkedIn, or Facebook. Go2Cab will help you quickly analyse text from hundreds of professionally accessed feeds to gain insight and make timely decisions about the following action.
- Video and Audio Analytics for any purpose, whether analysis of media clips posted on social media channels like Youtube or the analysis required for footage from a surveillance video or audio feed in industries like surveillance, safety, security or any operational environment where timely, relevant and accurate preventative measures are required.
- Mixed Analytics, where methods are used to analyse each type of analytics listed above, are different. Specific algorithms best suit the data and the intended purpose and use of the outcome. Go2Cab can analyse situations where your data is a mix of the categories listed above to be actionable to ensure that the analysis result is pragmatic.
- Predictive Analytics where, in most cases, the knowledge about something inadvertently went “wrong” is tolerable. However, undertaking predictive analytics is a “MUST HAVE” to achieve sustainable operational excellence. The challenge is to decide on the timing, the vital data, the type of analytics and the actions required to ensure operational preventative measures. Go2Cab uses best industry practices to ensure the analysis outcome is accurate, relevant, and timely before the “wrong” occurs.
- Forensic Analytics is where you need to analyse complex situational awareness promptly, where multiple factors are concurrently chaining in time, and you want to predict the following best [i.e., optimal] actions. Go2Cab can analyse complicated situations in near real-time, including factors like people’s behaviour. Such circumstances exist in everyday environments. For example, predicting and influencing the traveller’s behaviour is essential to achieve optimal traffic diversions in transport. Similar situations exist in planning events like sports, exhibitions, dignitary visits, security and surveillance. Predicting the most probable activity is critical to ensuring the optimal allocation, resource planning and execution, and the safety of all people involved. Decisions may have to change dynamically depending on real-time information and data feed.
Go2Cab recommends and is prepared to assist you pragmatically to instil the following actions into your operational environment to ensure that the outcome of the analytics has an immediate impact on the desired transformation in a manner that is, at all times, Accurate, Relevant and Timely at the lowest possible cost. These actions include:
Action 1: State Measurable and Testable Objectives
Start by identifying a few measurable and testable objectives that directly impact selected Customer needs. Validate such objectives against simple yet effective criteria to assess whether meeting them will delight the Customer or whether the Customer will consider your deliverables a “Have.”
It is constructive to note the keywords “measurable” and “testable”. The word “Measurable” incorporates essential decisions about what type of data you need, how often you need it, and margins for accuracy, which, in turn, will impact the mechanisms used to gather such data. There are several other factors that Go2Cab will help you identify and control using a sound scientific approach to ensure that such decisions are evidence-based.
On the other hand, the word “Testable” also incorporates several aspects that are well beyond how to test the expected outcome of the specific objective. Such aspects include specific measures relating to running the operation over time, the frequency and specificity of disseminating notifications, the timeliness and relevance of such notifications and any suggestive actions to retain, at all times, the optimal level of operation. To this end, Go2Cab uses a suite of commercial tools that will help you make decisions that will impact the solution you acquire and the configuration of the solution to avoid over-engineering and subsequent overspending.
Once you set the Customer objectives, you define the business objectives like that used when making decisions on Customer objectives. The recommendations made above should help you start thinking in the right mindset. It is essential to encourage “challenging” the decisions. It is equally important to verify and validate each objective iteratively. The terms “verify” and “validate” are two different things. “Verification” has more to do with making sure whether there is any apparent “missing” objective. In other words, you test the “Existence” or “Absence”. On the other hand, “validation” is more concerned with testing the outcome of the objective in “real-time” or “near-real-time”.
Go2Cab considers “validation” in the context of the Dynamics of the situations (i.e., the moving parts of the business). For example, one of the simple questions to consider in “validation” is whether the Customer notices a specific product feature or a particular aspect of the service. The other important aspect of “validation” is the delivery sequence of one or more service outcomes. The term “verification” tests the “existence” regardless of the “order” of the outcome. Validation ensures that the sequence of the delivery is timely and relevant. A simple example is to walk into an elevator (or a lift) where the doors close before you hear the announcement “Going up”. A valid challenge is to test the timeliness of such an announcement after closing the door, given that the elevator starts moving within half a second after the announcement.
Action 2: Avoid the Sole Reliance on Averaged Data, Simple Charts and Percentages
Assuming that the raw data is reasonably accurate, organisations tend to quickly present the “evidence” by averaging, calculating percentages and plotting various types of tally and charts. Such data may represent customer churn, proposed pricing for a product or service, time to respond to a call, time to discharge patients, number of incidents or time required to clear an incident, sales figures and pipeline over a prescribed period. If your industry is into health and food, such data may relate to tasting new products or general feedback about a new product during a pre-launch event. Suppose your industry is in Cloud-hosting, ISP, entertainment or utility services. In that case, such data may relate to downtime, churn, customer profiling, usage of cloud drive or memory space or some processors “provisioned” for the customer.
It is widespread to be influenced by the height of the bar in a bar chart or the size of the area of a pie chart. The first fact to consider is that the height of the bar (or the size of the region) may mislead you when making your next decision. The higher the bar (or more extensive area), for example, does not necessarily mean things are going very well or things depending on whether the bar is sales or better depending on several defects in a product.
A percentage of something is very similar. The small or large percentage does not mean things are very good or bad. For example, 80% of customers are happy, which does not mean that they have no complaints, nor does it mean that the “80” score is high enough. Even if you are in the 90% range, you are not better than a relatively low-ranked airline regarding their baggage handling service. Or, several downtimes per week if you are an internet service provider or have several thousand items returned every month because something went wrong. First, you must use the “right” evidence-based decision-making to set the “BEST” percent to meet over time. Such per cent, which GO2Cab calls the measure for CONSISTENCY, is calculated using methods other than simple Excel functions. When the raw data is analysed using the appropriate techniques, incorporating the pre-set objectives, you will find that the 80% was misleading.
Action 3: Simplistic Analysis
Most organisations need to be better advised when data gathering becomes the point of discussion. You can use simple charts, averaging data and percentages with CAUTION. Such simplistic reporting can (if not will) mislead you. Originations make decisions on the type of data and mechanisms to gather such data based on an already GENERALISED topic such as Stakeholder Management, Communications, Tracking and the like. Numerous details are obfuscated when most so-called experts (or advisors) are clueless in the domain. They need help dealing with complex interactions of parameters. It is also difficult for them to quickly prepare a few more PowerPoint slides and try to communicate to the C-Level, keeping it at a “high level”. Ironically, such advisors get away with such engagements as if the C-Level cannot understand the details. The trick is to present the proper measures in a simple, timely way, yet the measures incorporate the INFLUENCE of the business’s DETAILS, CONSISTENCY and OPERATIONAL DYNAMICS.
One of the consequences of doing it otherwise (e.g., the generalisation) is making decisions that there is, of course, engaging one of the survey providers to compile a survey and the advisor’s trust. The survey will be compiled using a set of questions that are already in the survey provider’s archives. In most cases, the questions are so generic that the context of the domain and the purpose of the data gathering is almost lost. To make things worse, the tally of respondents and percentages are reported where one can nearly guarantee that the next decision is sub-optimal, if not wrong. Most surveys deliver fake answers anyway. For example, relating to workload, you will need to find out if they are overloaded or underloaded. The answer is around 70% to 80% to avoid additional challenges where the respondent’s undesirable questions later have to be answered. Similar behaviour is proper when it comes to employee satisfaction. There is nothing wrong with conducting a survey. The trick is to design the study using an evidence-based approach, which Go2Cab calls Design Of Experiment (DOE). In this way, the data gathered serves a particular purpose by “Design”, and fake answers are significantly reduced or eliminated.
Action 4: Incorporate the moving parts of the Business
The advancements made in technology have enabled numerous organisations to make better decisions at a much lower cost. Those old enough to have used the early versions of Excel (known as Lotus 1-2-3) in the early ‘80s can tell the difference when using the recent versions of Excel (e.g. 2013 or 2016). That said, the advancements and rich features come at a cost. For example, consider the manufacturing technologies relating to fighter jets. The early versions of composite material used made it possible to camouflage fighter jets from enemy radar. Still, they made the jets more susceptible to radio interference because the original metal structure acted as a shield for avionics controlling various functionalities. Similarly, the recent versions of Excel have encouraged more people to export data off CRM or ERP tools and then author bespoke code to undertake analysis and generate reports.
One would have assumed that advancements in visualisation [business intelligence] tools like Tableau, Qlik, and a suite of tools from TIBCO (among many other equally respected vendors) would have significantly reduced the need to hire Data scientists to deliver and VB coders to bespoke Excel (VB) code.
Regardless of your destination, it is essential to consider two (amongst some other) factors. Firstly, Excel has no notion of units. You can manipulate the numbers with sophisticated formulas linking any cell to another. Excel is happy to produce a number unless basic mathematical rules are violated (e.g., divide by zero). The significance of this aspect exponentially increases with the complexity of the formula, where the tracking units get lost.
The other important aspect is that Excel has no notion of time (ie dynamics). It is easy to assign columns representing months or quarters or per cent in increments of 10 or 20 to cater for (say) growth or decline of something. Excel will not and cannot show the impact of business dynamics in time (i.e., how various factors, including customer behaviour (or any other moving part of the business), will interact throughout interest in configurable timeslots). The assigned columns where a formula calculates something are ONLY a snapshot of one specific point in time. Excel produces snapshots (like a still camera), whereas you need a “video” showing how things will shape up or down in time as you adjust various factors. It is the dynamics of the operational dynamics (i.e., the data, analysis, and reporting that are ACCURATE, RELEVANT, and TIMELY) that will make or break the organisation. When the C-Level or any MANAGEMENT level is aware of something going wrong, it is often too late to act.
Action 5: Avoid Premature Reporting
One natural yet disappointing outcome of generalising topics discussed during the information gathering and the relatively heavy reliance on highly aggregated data and simplistic analysis is the too-fast-too-soon decisions on tools. The other reason for making premature decisions when selecting and configuring one or more tools is to quickly please senior-level managers and executives. Despite the rigorous methods used to choose the tools, trials and proofs-of-concept, the reality is that some vendors will convince the organisation that they have the best tool, and the “out-of-the-box” slogan gets things going for a while until the honeymoon is over. Such an unfortunate outcome is not the vendor’s fault. The root cause for such faulty outcomes is the mechanism to gather and process the information leading to the generalisation and the way data were compiled, analysed and presented, missing the first crucial steps as explained above. Since almost all tools have some form of reporting capabilities (now sold more like “ANALYTICS”), within 6 to 12 months, the organisation would have customised numerous reports. After a typical period of one year, the organisation realises that nothing has changed regarding efficiency, an increase in profits, or better serving the Customer. Further, the organisation realised that the so-called “Analytics” was no more than simple averaging, per cent and simplistic calculations but fairly robust visualisation of data with little or no “ANALYTICS”.
Examples of poor decisions in the context of Action 5 can fill a dictionary-thick book. Reports and dashboards show daily or monthly satisfaction of call centres where the gauge never moves a few ticks below the max. The reason is mainly the generalisation and simplistic calculations; the score for “Answering the Customer call is 4.9, 4.92 or 4. 96 out of a total scale of 5 reported every week or every month. No one questions that the score will NOT be much different when one averages 3000, 2700, or 2800 data points (e.g., time to respond to a Customer’s call). Further, no one questions what the number means from an operational dynamics standpoint.
In most cases, the average numbers are measured from when the phone rings to when an operator lifts that handset (i.e. the call centre staff say: “Hi, my name is Joe, how can I help?”). One wonders if the executives know the dashboard does not meet customer needs. Such misleading reporting is “ALL GOOD” since everything is always GREEN, which Go2Cab calls “Watermelon Reporting; “all green from the outside but red inside!” A member of staff answering the call in 10 seconds has very little to do with whether the call centre has addressed the customer’s needs. The Customer may have been placed on “Hold” for another 5 to 10 minutes per call, and it possibly took ten more calls over 3 days before the call to FULFIL the Customer’s centre managed need. The same is true when dealing with time to clear incidents reported per unit of time (be it a week, month or quarter), sales growth, or customer churn.