This is part 11 of a 12-part series documenting what I learned from the “Technovate Thinking” course at business school.
The pre-session assignment for the final class was: Following the Technovate Thinking problem-solving process, use free tools to design and implement a solution for a real problem in your own business. This was the true capstone — pouring in everything I’d learned from Sessions 1 through 5 about frameworks and tool usage.
Honestly, when I first heard the assignment description, I thought, “That’s heavy.” But at the same time, it was a chance to consolidate everything I’d learned into a single deliverable. I decided to pick a theme based on something I was genuinely struggling with at work.
- 1 The Problem I Wanted to Solve: Failing to Notice When Key Relationships Deteriorate
- 2 The Starting Point: Cialdini’s “Six Principles of Influence”
- 3 What Is “Virtue Accumulation”? — Quantifying Everyday Actions
- 4 Design: Mapping Relationship States with a Transition Diagram
- 5 Five Data Categories: What to Collect and How
- 6 Implementation: Network Visualization with Gephi
- 7 Three Benefits This System Delivers
- 8 Four Challenges I Honestly Acknowledged
- 9 The Boundary Between Technology’s Limits and Human Judgment
- 10 Retrospective: Everything from Sessions 1-5 Was Connected
- 11 Next Time (Series Finale): Reflecting on All Six Sessions, and What Comes Next
- 12 Books Referenced in This Article
The Problem I Wanted to Solve: Failing to Notice When Key Relationships Deteriorate
The theme I chose was relationship management with client key persons.
Working in IT infrastructure, there’s always someone on the client side whose judgment drives the deal. When the relationship with that person is healthy, the project runs smoothly. Meetings have a positive atmosphere, information flows naturally, and when issues arise, both sides approach them with a “let’s solve this together” mindset.
But when something causes that relationship to sour, the entire engagement can stall or the client may switch to a competitor. And more often than not, by the time you notice the deterioration, it’s already too late.
The warning signs are there. Email replies start taking more than three days. They speak less during meetings. Conversations that used to include casual chat become strictly business. Individual account managers “sense” these subtle shifts. But there’s no mechanism for the team as a whole to capture and act on them early. Everyone relies on individual “gut feel” — that was the problem I wanted to solve.
The issue of over-reliance on individuals is severe. When the account manager transfers or leaves, that “gut feel” walks out the door with them. The new person starts from a blank slate, and during the ramp-up, the relationship risks deteriorating further.
The Starting Point: Cialdini’s “Six Principles of Influence”
To give structure to this vague problem, I searched for a framework. What I ultimately chose was Robert Cialdini’s “Six Principles of Influence.”
Cialdini identifies six psychological mechanisms at play when humans are influenced by others. I figured this could work as a framework for decomposing the “quality” of a business relationship. Let me walk through each principle with concrete business examples.
Reciprocity
When someone does something for us, we feel compelled to return the favor. In a business context, proactively sharing valuable information, lending a hand when they’re in trouble, going the extra mile on small things — these actions accumulate as “credits.” Conversely, a relationship where one side always takes without giving gradually increases the other party’s resentment.
Commitment & Consistency
People strive to behave consistently with commitments they’ve made. For example, if a key person has declared internally, “We’re going with Company A on this,” they’ll want to maintain consistency with that statement. Whether this “commitment strength” is being maintained is a crucial indicator of relationship stability.
Social Proof
Whether the people around someone support a decision influences their judgment. If the key person’s surrounding team members are favorable, the relationship stays stable. When negative voices start emerging from the periphery, the key person’s own attitude is likely to shift as well. Relationships need to be viewed not just one-on-one, but including the surrounding network.
Liking
Simply put: “Do they like you?” Shared hobbies, values, and similarities foster liking. In business, the frequency and content of casual conversation serve as indicators. If someone who used to chat enthusiastically about golf now only sends transactional messages — that’s a sign that liking has decreased.
Authority
Influence based on expertise or position. Is the key person accepting your proposals as “expert opinion”? If their stance has shifted from “You clearly know this area — I’ll leave it to you” to “Let’s hear what other vendors think too,” that’s a sign your perceived authority is declining.
Scarcity
People value things that are hard to obtain. Is the information or expertise you provide seen as “available only from you,” or as “easily replaceable”? Changes in this perception are closely tied to competitor activity.
By scoring relationship quality across these six dimensions, instead of “the relationship vaguely feels good/bad,” you can structurally understand which aspects are strong and which are weak. This is where the concept of “virtue accumulation” was born.
What Is “Virtue Accumulation”? — Quantifying Everyday Actions
“Virtue accumulation” is a system that quantifies small everyday actions as “virtue” points and tracks them over time.
For example: sharing a valuable industry insight via email earns “Reciprocity +2.” Allocating resources to help with a difficult project earns “Reciprocity +5, Liking +3.” Presenting a new proposal in a regular meeting earns “Authority +2, Scarcity +1.” Expressing gratitude earns “Liking +1.”
Conversely, missing a promised deadline results in “Commitment & Consistency -3.” Extended periods without contact result in “Liking -2/week.”
As these scores accumulate, a “virtue balance” becomes visible. As long as the balance is healthy, the relationship is stable. But when it drops below a certain threshold — that becomes the trigger for an early warning alert.
The key insight is that this system captures not just major events (incidents or contract renewals) but the accumulation of small everyday actions as numerical data. I chose the word “virtue” precisely because it captures the essence of this gradual accumulation.
Design: Mapping Relationship States with a Transition Diagram
Using the Transition Diagram approach from Session 2, I defined key person relationships as four states.
Healthy: High virtue balance with well-balanced scores across all six principles. Active information exchange, and new project inquiries come in naturally.
Caution: Some scores have started to decline. Communication frequency may be maintained, but meeting participation has dropped, or response times have slowed. Intervention at this stage can restore the relationship to “Healthy.”
At Risk: Multiple scores have fallen below thresholds, with clear behavioral changes. Meeting cancellations increase, competitor names start coming up. Recovery from here requires significant effort.
Lost: The relationship has effectively ended. The engagement was terminated, you were removed from the account, or communication has completely ceased. At this point, recovery through normal means is extremely difficult.
I defined transition conditions between states using a combination of score thresholds and behavioral patterns. For example, the “Healthy to Caution” trigger is “virtue balance drops below 70%” combined with “response speed deteriorates by 2x or more.” The “Caution to Healthy” recovery condition is “virtue balance increases for two consecutive weeks” combined with “meeting attendance above 90%.”
The process of drawing this Transition Diagram was exactly the “envision the desired state” exercise we learned in Session 2.
Five Data Categories: What to Collect and How
I designed the system to collect source data across five categories for scoring.
① Communication Frequency
Number and direction (outbound/inbound) of emails, chats, and calls. A decline in communications initiated by the other party is the most obvious early warning sign of relationship deterioration. This can be automatically aggregated from email and chat tool logs.
② Response Speed
Average time from sending an email to receiving a reply. A sudden slowdown signals that something has changed. However, the other party’s workload and project circumstances need to be factored in. I designed the system to compare the “two-week rolling average” against the “three-month average” to determine whether the value is anomalous.
③ Meeting Attendance & Engagement
Attendance rate at regular meetings, frequency and quality of contributions. “Attends but says nothing” is an important signal for transitioning to “Caution.” This category is difficult to capture automatically and relies primarily on manual input.
④ Project Progress
Whether active projects are on track. Frequency of delays and scope changes. It’s not a simple “smooth project = good relationship” correlation, but project trouble is a common trigger for relationship deterioration, so I included it in the monitoring scope.
⑤ “Virtue Accumulation” Score
A composite score derived from categories ① through ④, plus subjective assessments based on Cialdini’s six principles. The design assumes team members spend about five minutes per week entering data, prioritizing “rough but sustainable” over “perfect data.”
Implementation: Network Visualization with Gephi
For implementation, I used Gephi — an open-source network visualization tool that renders graphs using nodes (points) and edges (lines).
Specifically, nodes represent “people” and edges represent “relationships.” I placed our team members and the client’s key persons and surrounding stakeholders as nodes, then varied edge thickness and color based on communication frequency and scores.
For example, a “Healthy” edge is a thick green line, “Caution” is yellow, “At Risk” is a thin red line, and “Lost” is a dotted line. This way, when you look at the entire network at a glance, you can visually spot where things are weakening.
Additionally, by graphing score trends over time, you can see things like “This was a thick green line last month, but now it’s turned yellow” — meaning “the relationship with this person is thinning” becomes visually apparent. This kind of insight simply isn’t available from a spreadsheet of numbers alone.
I chose Gephi for three reasons: it’s free; it’s specialized for network visualization and performs well even with many nodes; and CSV import is straightforward, making it accessible for teams managing data in Excel. This was a practical application of the “free tool selection criteria” from Session 5.
Three Benefits This System Delivers
I organized the system’s benefits into three categories.
Benefit ①: Eliminating Dependence on Individuals
Relationship management that previously depended on one person’s “gut feel” becomes data the entire team can share. Even when account managers change, reviewing the historical score trends reveals the relationship’s history. The quality of handoffs changes fundamentally.
Benefit ②: Early Detection
By regularly monitoring numerical trends, you can catch issues at the “not critical yet, but heading in a dangerous direction” stage. What used to be “I have a vague sense something’s off” gets replaced by data telling you “scores are down 20% compared to last month.” That 20% difference is what separates being able to act from being too late.
Benefit ③: Reducing Emotional Burden
This one is surprisingly significant. The feeling of “something’s awkward” accumulates as unspoken stress for the account manager. When data allows you to articulate it as “The Liking score is declining. Let’s increase Reciprocity-focused actions” — it transforms from an emotional problem into a problem-solving exercise. It becomes easier to figure out what to do, and easier to escalate to a manager.
Four Challenges I Honestly Acknowledged
I didn’t stop at benefits alone, because the “verify and improve” mindset reinforced throughout Technovate Thinking wouldn’t let me. I confronted the challenges head-on.
Challenge ①: Japanese Sentiment Analysis Accuracy
AI that reads emotional tone from email and chat text is reasonably accurate in English. But Japanese relies heavily on honorifics and indirect expressions, making it difficult for AI to determine whether “We would appreciate your kind consideration” is genuinely positive or a veiled expression of dissatisfaction. Until accuracy improves, the system design must avoid over-reliance on text analysis.
Challenge ②: Data Entry Burden
The act of recording daily actions is itself a burden for people on the ground. Even though emails and calendars can be auto-aggregated, subjective data like “How did the other party react in today’s meeting?” requires manual input. The most realistic risk is that the system becomes a chore and falls into disuse. I designed it around “five minutes per week” specifically to minimize this risk.
Challenge ③: Psychological Resistance to Quantifying Human Relationships
This was the most deep-rooted challenge. “Managing virtue with scores — isn’t that too calculating?” Some people will inevitably feel this way, and that instinct isn’t wrong. There are aspects of human relationships that can’t be quantified. Forcing them into numbers might even distort the relationship itself.
My answer to this challenge is to position the system not as a pursuit of perfect scoring, but as a trigger for awareness. When a score drops, it’s a prompt to ask, “Why?” The score itself isn’t the answer — its fluctuation is a catalyst for conversation.
Challenge ④: Initial Investment Cost
Gephi itself is free, but preparing the data, building the input workflow, and educating the team all require non-trivial time and effort. “Start small with one key person” is probably the most realistic approach.
The Boundary Between Technology’s Limits and Human Judgment
The biggest takeaway from this assignment was that “resistance to quantifying human relationships” is a human problem that technology cannot solve.
Technology can collect data, detect patterns, and raise alerts. But deciding “how to respond to this change in numbers” and “whether the situation is truly as the score suggests” — those judgments can only be made by humans.
The philosophy underlying Technovate Thinking is: “It’s humans who wield technology.” Through this final assignment, I internalized that philosophy at a visceral level. The ability to draw the line between what technology “can do” and what it “cannot do” — that’s the very capability this course was designed to cultivate.
Retrospective: Everything from Sessions 1-5 Was Connected
Looking back, the final assignment was structured to use everything I’d learned from Sessions 1 through 5.
Session 1’s “problem-solving in the Technovate era” taught me to “define the problem to solve.” I transformed the vague notion of “relationship management isn’t working” into the specific problem of “a structure dependent on individual gut feel.”
Session 2’s Transition Diagram taught me to “envision the desired state.” By diagramming four states and transition conditions, the overall system design came into focus.
Sessions 3 and 4 on algorithms and scoring taught me “how to design the solution.” The process of converting Cialdini’s six principles into scoring axes was algorithm design in its purest form.
Session 5’s tool utilization taught me to “get hands-on with implementation.” Visualizing the network in Gephi and actually feeding in data revealed the weak spots in my design.
And finally, organizing the benefits and challenges to make a judgment. Having completed this entire process solo was an embodiment of Technovate Thinking’s core principle: “Logical design is the human’s job.”
Humans use the tools. But humans also decide what the tools should do. Becoming a business leader who can do both — that, I believe, was the ultimate message of this course.
Next Time (Series Finale): Reflecting on All Six Sessions, and What Comes Next
The next article is the final one in this series. I’ll write about what changed over all six sessions, what I’ve brought back to my day-to-day work, and what I plan to learn going forward.
Books Referenced in This Article
