XIX. You need a MAP

Divide and conquer by structuring your research space for Metrics, Architectures, and Problems
Dr. Albert Künstler

ow that you feel comfortable with uncertainty, it is a good time to lay them out on a MAP. You will have hundreds of unknowns, your uncertainty tickets, and your future work items, so it makes sense to keep them sorted, arranged, and structured instead of having a huge pile or a long list of random tickets of all kinds and sizes.

Brave explorers make notes as they travel and record on the map landmarks and rivers, jungles and deserts, lakes and hills, and towers and bridges, as well as discovered ruins of ancient civilizations. They also make assumptions about the unexplored areas and continuously reveal the map as they travel.

The same should be done even if you are not a sailor or a traveler, but you work on solving business problems. In your case, it will not be a mysterious island, but several visual structures that would benefit from a virtual whiteboard or a physical investigation board in your office.

Tool: MAP: Metrics, Architectures, Problems

Principle: FIGARO: Filling identified gaps, allowing reasonable overlaps — for dynamic disaggregation of yet uncertain research space

Best for: Disaggregating and visually presenting areas of your future research

Required time: 60–90 minutes per map

Main result: A shared “big picture” and a map of your studies that will reveal as you explore the areas.

There are three structures that are most valuable that will collectively define the MAP of your research. Most likely, you need all three of them in your quest.

Metrics — A metric tree (or several trees) will be an important part of your future studies. Quantified metrics are essential for measuring success, estimating impact, evaluating the results of experiments, and comparing discovered alternatives. At the beginning of your Quest, you usually do not have a full understanding of success criteria, but only vague apprehension of what is Good and what is Bad (and something that is still Unknown). This is okay. Your future efforts will be aimed at disaggregating the metric tree, exploring the relations between business KPIs, establishing new measurement procedures when needed, and connecting them to your north star metric.

Architectures — The architecture diagram (or several) will help you disaggregate and represent systems and processes. They are essential when you are trying to build, optimize, or re-engineer something. There are several types of architectures that will be discussed in the next chapter. One of such important structures is the DTBM Logic that represents the steps and sequences of how the quest givers expect their Question to be solved. Having architectures as a part of your MAP helps you to define the research space and organize and pin uncertainty tickets on this board.

Problems — Disaggregated problems is another space of your future research. By structuring this space, you will be able to attack a big problem that may be too strong to be solved completely. However, conducting a root-cause analysis and presenting it as a combination of smaller partial problems will help you to find a set of economically viable partial solutions to the highest impact sub-problems and thus bring the quest givers what they need.

How these 3 MAPs will aid you in your journey

As we well know, your quest givers need to make an important decision — they have a question that must be answered or a problem that must be solved. Often that means you need to explore, fix, or build something. In most cases this “something” is either a process or a system.

Thus, you are expected to

  1. Understand the existing system or process (AS IS)
  2. Identify the missing or improper parts that cause problems
  3. Combine alternative solutions and suggest new experimental designs
  4. Test alternatives and solidify the improved architecture (TO BE)
  5. Build a new process or system that works better

Steps one through four are pure Discovery, and step five is both Discovery and Delivery.

For the first step you absolutely need a clear disaggregated picture of the Architecture of the existing system or process without missing important details.

For the second step, you are expected to analyze the sources of Problemsand disaggregate them into partial subproblems and decide which of those will you address first. Exploring the problems you also come across possible solution ideas.

For the third step, you begin by combining these ideas, re-engineering the system or the process, and then prototyping and populating the redesigned Architecture with the solution ideas.

Step four, testing, will require well established and quantified Metrics, so you’ll be able to compare the outputs and outcomes of the experiments and solidify the new improved Architecture.

The last step (Implementation) is not always expected from you on your quest. Sometimes changes are too expensive or complex and will be carried out by the quest givers and their team. In some other situations, this is a continuous discovery process, so building, changing, and experimenting work in a loop. As long as we discuss generalized cases, we will assume your crew is involved in this building activity as well.

Disaggregation techniques

Your MAPs (Metrics, Architectures, and Problems) should be broken down into smaller Areas. These research areas are pieces of the bigger map that may become targets of a more focused exploration. For example, areas may be associated with the individual components of the system, steps of the process, business KPIs, partial problems, research topics, alternatives, or other particular pieces of work.

You typically want these areas to be small enough that one researcher can study them completely, acquire a deeper understanding, and communicate back to the team. That way, knowledge about the area becomes shared and this piece of MAP becomes collectively explored. At the same time, a single area should be at that level of granularity to contain several interconnected Uncertainties, so that the researcher who has spent some time exploring it will find solid answers, without the need of expanding too far to adjacent areas.

Although there is no one best way of disaggregation and grouping and the same territory may be divided into parts in different reasonable ways, there is a good principle that is quite popular among architects and consultants. It is known as the No Gaps — No Overlaps principle or MECE: Mutually exclusive and collectively exhaustive parts.

The idea is simple: there should be No Gaps — the areas you create should cover the entire map, so no piece, subproblem, component or metric is left unattended, they are collectively exhaustive. There should also be No Overlaps — the areas should not intersect, so the same problem or component should not be part of two or more areas, and you don’t need to do the same work twice.

Disaggregation should not be considered as a one-time event. When we are trying to divide the uncharted land, we have a very limited apprehension of the unknown territories, but don’t hesitate to make plans — just be prepared for reconsideration. Some areas may appear to be larger than you expected, and you will split them into two, three, or even hundreds (such “cambrian explosions” will be discussed later) of areas.

Some inconveniences may also arise down the road with the No Overlapsprinciple. During the exploration phase you may find an unexpected relation or dependency between two areas so that neither of them can be resolved separately — for example, two problems with the same root, or several components of the system that have shared function. In such an occasion, you may stick with the No Overlaps principle and try to decouple the shared space into a new area, so the two original areas will shrink a little and their shared piece will gain a separate status, but still remain mutually dependent. Or, alternatively, you may avoid excessive purification and rely on the more flexible Filling Identified Gaps, Allowing Reasonable Overlaps principle.

FIGARO: Filling Identified Gaps, Allowing Reasonable Overlaps

FIGARO is good for the exploration process in its dynamics, when you have a “sort of” MECE at every moment of time. However, you are tolerant to emerging areas that are recently discovered and vague and thus not yet collectively exhaustive, and tolerant to some interdependent areas that are not completely mutually exclusive. You consciously keep these inconsistencies as-is, because regrouping too often is an expensive operation and under the current level of uncertainty may be counterproductive. After some period of time, as you obtain more clarity in these areas, you may reorganize and reattribute the uncharted areas, but by that time you’ll have more discoveries and more inconsistencies somewhere else. So keep moving. FIGARO here, FIGARO there — that is the way.

Planning techniques

Once you have more or less defined structure areas, you have a wide space for your research and thus there is too much to be done. Here comes the question of prioritization. Which areas deserve the most attention? How much effort should be made here and there? We will discuss prioritization methods in a separate chapter, but for now, let’s remind ourselves of the Level Up principle. You should not only prioritize some areas over the others, but also define the Level of desired elaboration. For example, you may spend all your resources and invest in a deep research of one particular area, or you may spread resources thick and scout around doing quick research in several areas.

Start with quick scouting, get some findings, then dig deeper into the most promising areas.

While the general rule is “Start with quick scouting, get some findings, and dig deeper into the most promising areas,” the real world is more complex than that. Every architecture or type of problem has its own recommended leveling strategy.

In general, we will follow the Zero-to-Hero prioritization principle that advises to develop the pieces of the MAP in a particular sequence. Although the research is an everlasting continuous process, our quest should have several phases, when some parts of the MAP become more important than the others and require more targeted effort. These target areas on each phase should accumulate most attention and upgrade following the Fidelity vs Quality curve from Zero to Placeholder then to Quick & Dirty then to Good Enough, then to State of the Art level, from being an “Ugly Baby” through “Dexterous Apprentice” quadrant up to the “Mighty Hero”. Some parts will move faster than others, some will stop at a certain level and wait others to upgrade.

This coordinated upgrade strategy for each type of MAP requires a separate leveling guide. In the next chapters we will discuss the best practices and general guides for all three MAPs in more detail.