Skip to content
back to the episode
Episode summary2022Fellow

Technology Strategy Patterns — Episode 4

The fourth Technology Strategy Patterns episode covers a single chapter: how to communicate a strategy once you know what to change. Alexander, Vova and guest Evgeny Peshkov, who builds community and works on strategy, walk through seven patterns: the 30-second answer, the rented brain, Aristotle's rhetoric, the fait accompli, dramatic structure, deconstruction and the business machine.

Code of Architecture · Book Club9 min read

This summary is shorter than the full episode; examples and assessments reflect the participants’ experience and views.

The main thread of the material
01

Thirty seconds and the rented brain

The episode opens with a question from the comments under Vlad Khononov's book, also put to the author himself: are there successful DDD examples in practice? For Evgeny it is not a ready-made tool but a set of practices, some of which will certainly work; Alexander agrees. Evgeny praises the chapter and would move it to the front of the book: you can devise a brilliant strategy, but if you cannot sell it the work is wasted. Alexander read the book in 2019 before a big challenge and felt a clear before-and-after difference. Vova leads the first pattern: answer a stakeholder with at most three points, each with a reason (autotests shorten regression, speed up delivery and are fashionable), then stop in time rather than spray terminology. The boss is human too: eighteen meetings today, and your window is his dash to the toilet.

Evgeny calls the startup-pitch comparison spot-on; the participants add that in remote work the thirty seconds are a window for reading the answer, not writing it. Rented Brain is about hired consultants, called in when a company lacks experience of its own. The author dislikes them: having solved one problem, a consultant creates a bigger one and gladly takes that on too, like a garage that finds a timing belt after every repair. The valuable part is the mental shift: imagine you are the smartest person in the room, drop your career worries and say what has to be said. Inside a company everyone is doing great, and a sensible executive values the one person who named the problem. Alexander recalls a book titled roughly Sorry, I Wrecked Your Company, by a consultant who left her consulting firm to start her own, and the story of the big three — BCG, Bain and McKinsey — grown out of the shift from factory owners to hired managers. His metaphor: the consultants set up your pieces by move six of the opening, but you have no plan for the game. An anti-pattern from the participants: a paid consultant is trusted more than your own people.

02

Aristotle's rhetoric and logical fallacies

Vova takes Aristotle: in the Rhetoric a convincing position rests on logos, ethos and pathos, and all three are needed — drop one and the argument sags. Logos is reasoning the audience walks through with you until it reaches the hypotheses itself. Ethos is authority: everyone knows Vasya, and Vasya pulled it off, although perhaps by accident or because of a company merger. Pathos is emotion: if we do not act, production goes down and we lose part of the business; if we do, paradise awaits. Do not overdo it: stakeholders are not stupid and spot a show instead of useful information. Alexander warns against extrapolating authority — Vasya wrote a book on performance, which does not make him an investment expert — and against the favourite argument that Google, Apple and Amazon do it. To logos he adds progress metrics announced in advance: otherwise, like the Texas sharpshooter, you will later pick the numbers that happened to improve.

Alexander opens the catalogue of fallacies with a nod to Thinking, Fast and Slow: the fast system of thinking misfires, and misfires turn into crooked arguments. Ad hominem: Vasya never built a high-load system, so ignore him on the crypto exchange architecture. He separates stereotype from statistics on the spot: twenty rejected candidates from the green company are not yet a representative sample. Affirming the consequent: an alert that fired does not prove the load was the thousand RPS at which the server should fall over. Appeal to authority turns into cargo cult: build an airport and the goods will drop from the sky. Clever words collapse when you ask the speaker to go deeper. Hasty generalisation: a good October means we perform all year, when in fact nobody was ill. Petitio principii is a circular argument, and the last one draws conclusions from a coincidence in time: A and B may both depend on a hidden C. Still, the event before an outage is the prime suspect, and the author should have mentioned Occam's razor.

03

From fait accompli to the business machine

Fait Accompli is the meat of the chapter: the idea has to become reality before you ask for a budget. The anti-pattern is a cold audience; in the participants' experience such a meeting turns into nitpicking, or silence with no feedback. Hence the corridor warm-up and the stakeholder hunt: there is always at least one more stakeholder than you know of, and one more requirement than they name. Everyone tries the new world on for size, and the first reaction is rejection; talk to those who will resist hardest and weave their ideas into the deck, so the room is not against you but against you and Vasya. One host recalls a manager who made sure people felt the decision was their own: without that skill you cannot grow past a certain level. Alexander extended Dramatic Structure with the monomyth and The Hero with a Thousand Faces: people think in stories. Vova lays out the Hollywood script — status quo, an incident such as hardware deliveries stopping, the hero's plan, shock and awe, victory over Goliath — with metrics and a real plan, such as renting hardware.

The pattern doubles as an indicator: if the shock and awe will not come, there is probably no problem and you are simply holding a hammer. Evgeny does not see Deconstruction as philosophical: problems come from the things we are sure cannot happen. You optimise regression, a queue grows in the pipeline, and critical services deploy slower than your own toy; everything is a trade-off, so you take an assumption and its opposite and catalogue the hypotheses. Freud's line is worth keeping: the patient does not remember what he forgot yet repeats it — that is about local optima and crutches, while a knowledge worker's job is to eliminate the class of problem. Vova unpacks the business machine: actions and roles, deliverables, output for what was shipped and outcome for why, departments such as his own business lending unit, business units with a P&L, and the company as a legal fiction above them. Fifteen steps run from a vision to metrics, templates and hot spots, and the hard part is implementing rather than drawing. One chapter took nearly a hundred minutes; next Thursday brings the finale, with a picture where every pattern is joined by arrows.

Takeaways

What to take away

  1. 01Answer a manager with at most three justified points and then pause; in chat the thirty seconds are for reading the answer, not for writing it.
  2. 02The rented-brain stance means saying what has to be said; the danger is not consultants but hiring them thoughtlessly, when the company takes nothing over from them.
  3. 03Logos, ethos and pathos work only together, and logos needs metrics announced in advance; run your arguments past the fallacy catalogue, from ad hominem to correlation-as-cause.
  4. 04Big changes are sold before the meeting: find the stakeholders and weave in their ideas; if you cannot describe the shock and awe, there may be no problem at all.

Sources