It’s relatively easy to get caught up in a numerical frenzy that distances us from or distorts the essential process of (Extracting) Data -> (Transforming into) Information -> (Delivering) Knowledge, which can be overlooked or treated elliptically even though the “core” (let’s not forget) will always be the generation of value through the delivery of Knowledge.
As the iconic Marvel factory would say: with great power comes great responsibility. Leaving aside the ‘hard skills’ part, which are already covered (and very well) in multiple Masters and Postgraduates, let’s focus on the more ‘soft’ part and show 5 ‘dos’ and ‘don’ts’ to consider in profiles related to Digital Analytics. There will always be more, and those listed here can be subject to discussion. Let’s go.
The 5 “do’s”
1. Create a Data Culture Within the Organization
Sounds exciting, but there’s nothing more difficult than changing ingrained dynamics and trying to develop new ones: “old habits die hard.” Creating a data culture is a ‘step beyond’ that is impossible to achieve if the organization doesn’t firmly BELIEVE in it. The Digital Analyst must act as a catalyst, accepting that they can hardly reach the goal alone.
Therefore, it’s about overcoming the “functional/departmental area” vision so that analytics permeates the organization, from top to bottom but also horizontally.
2. Understand and Contrast the Marketing Funnel Within the Analytics Process
It’s necessary to periodically step back from the data and contextualize it within a marketing process. All users go through the stages of “Need -> Information Search -> Evaluation of Alternatives -> Purchase -> Post-Purchase”. It’s necessary to compare what happens in our digital “upper/mid/lower funnel” with the stages of the more dogmatic marketing funnel.
3. To Formulate the Appropriate Hypotheses
The formulation of hypotheses is related to a concept that we should, if you’ll allow me the hyperbole, have branded into our minds like cattle: ‘data as a means to an end’. We must formulate the hypotheses to be tested and design the analytical roadmap to reach a conclusion. The hypotheses must be purely ‘business-oriented’. That is, they must address very specific problems and situations. Starting from a correct fixation of the hypothesis to be tested (for example: I am losing users at this stage of the pre-checkout process and it may be due to factor X), we develop the analysis design that should allow us to validate/discard said hypothesis with a degree of confidence of X%.
4. To Understand the 4 Perspectives of the Balanced Scorecard
What, Why, When, and For Whom. Simple as that.
- What a Balanced Scorecard is and what it isn’t: it should be a representation of the main key performance indicators (KPIs) in numerical, graphical, and written form, and combine a static and dynamic view. It’s not a tool exclusive to top/middle management. Didn’t we talk about data culture earlier?
- Why it’s necessary: it should allow for a global view and align the organization in terms of objectives. It should allow for knowledge sharing and be an element that “drives” people and the organization in a specific direction. Quite a lot, actually.
- When: whether they are ad-hoc/one-time or recurring – and in this case, what the frequency should be depending on the content, its urgency, and recipients, among other factors.
- For Whom: it’s essential to know and understand who the final recipient of the Balanced Scorecard is. Before attending any preliminary meeting, it’s necessary to always request the list of attendees and track them (professionally speaking, of course): who they are, what their background is, what they expect to obtain, and what their level of decision-making power is.
5. Storytelling
We must be able to break down all that knowledge in a structured and scripted way. Storytelling is partly responsible for triggering the actions we want to happen. It is as important to deconstruct the entire process of “Data Extraction -> Analysis/Transformation into Information -> Knowledge Delivery” as it is to create an understandable discourse.
The Digital Analyst must be a person capable of leading meetings, adjusting the discourse to the interlocutors. There is nothing worse than “losing the troops” due to an unstructured or unadapted discourse.
The 5 “dont’s”
1. D.T. or “Data Thrower Syndrome”
From data thrower to value generation. The Digital Analyst isn’t there to simply ‘throw’ data around. It’s easy to fall into that dynamic, subtly and sometimes imperceptibly. Having powerful analytics tools, high traffic, and multiple acquisition sources can draw us into that vortex.
There’s no magic cure. When this happens, it’s necessary to overcome these phases of excessive exploitation and focus on: extracting, analyzing, and synthesizing.
2. Not Knowing How (and When) to Say “No”
More and more data is requested from various departments and functional areas, the request is collected, exploited, and shared… We must learn to say “No”. Humility and assertiveness, two sides of the same coin. Making a selection of what is important at the moment is an essential function. What is analyzed is just as important as what we discard or put on hold.
The ‘no’ must be argued and should shift from a request/imposition model to a conviction-based model.
3. The 4 “Demons” in a Dashboard
Just as there are ‘dos’ in a Dashboard, there are also ‘don’ts’. A Dashboard is a tool that should allow access to Information (not raw Data). Although there are different types of errors, below I summarize what I consider to be 4 major mistakes to avoid:
- a) Imbalance between Static vs. Dynamic Vision: both must be reflected to understand what is happening within a broader context that allows for the display of trends.
- b) Interpretations vs. Hypotheses: understanding and interpreting unequivocally. Where does insight end and hypothesis begin: a mistake to avoid is being “in limbo” and elevating said “insight” to the category of “hypothesis” when a previous analysis would have led us to discard it.
- c) Too much information: what we could colloquially call “going overboard”. We have a lot of data, but not all of it can be transformed into information on its own, nor is it necessary to dump “all” the available information.
- d) And the most serious: “non-actionable” information. It already happened to Mickey Rourke in “Angel Heart” with an unhappy ending (sorry for the spoiler): How terrible is wisdom that does not bring benefits. The “Dashboard” must be a catalyst for action. It must “trigger” the teams, yes, but in the right direction.
4. Lack of Business Acumen
The Digital Analyst must be ‘business oriented’. They say, ‘It takes two to tango’. I would add, ‘It takes two to kiss’. There must be a two-way approach: from the analyst to the business, and from the analysis requesters to the analyst. If there’s a Data Thrower Syndrome, there’s also an Analyst in a Corner Syndrome.
5. Failure to Grasp the Organizational Context
The Digital Analyst must accurately assess the stage at which the Organization is, its stakeholders, and the ‘resistance to change’ (if it exists). Many organizations have yet to complete their digital transformation. We can think of ‘diplodocus’ companies (understood in terms of both size and longevity, without any negative connotation) and/or those operating in sectors that we consider more traditional. Digital transformation, understood as the replacement of processes, positions/people, and organizational models, has not yet been completed, and yet we are already asking them to execute a new layer of transformation based on a data culture. Not managing the ‘timing’ correctly is a mistake that can engulf not only the Analyst but the entire team.
