Takeaways: Good Strategy / Bad Strategy (Rumelt), Part 7

Chain-link systems and Excellence (41% completed)

The focus was well-received! Going to keep these more topical as a result. Today’s was an incremental and beautiful articulation of the double-edged sword of chain-link systems.

⏮  Start at Part 1 | ⏪ Back to Part 6


New chapter! Chain-link Systems, in Part II: Sources of Power. “A system has chain-link logic,” says Rumelt, “when its performance is a limited by its weakest sub-unit or ‘link’.”

  • Takeaway #1: “When there is a weak link, a chain is not made stronger by strengthening the other links”

If you’re like me, you’ve done some work in startup-land, which has some chain-link systems at play. Product management is an easy one to reason about. A well-managed product management function will deliver delightful, hard-to-copy, margin-increasing features.

If they’re not doing that, then what is sales selling? what is marketing talking about? What is engineering building? Say, if you double-down on your investment in sales to compensate for middling product performance, does that help? I think this is chain-link logic in action.

  • Takeaway #2: “Quality matters when quantity is an inadequate substitute”

Continuing on the topic of product management, if PM leadership isn’t effective, would hiring 3 more product managers help? This would be a different situation if we were digging a hole or doing data entry. Quality leadership > quantity of contributors.

  • Takeaway #3: “When assessing potential, one should identify the limiting factors.”

Rumelt’s example in this section was a home purchase. House next to a highway? Not going to change that, regardless of how much marble is in your bathrooms.

⭐ My mind immediately went to the hiring process. The way I interview engineers, I accumulate reasons to (and reason not to) hire someone, which eventually becomes a case to take some specific action (sometimes sooner rather than later ☠).

⭐ Are they fresh out of school? Do they have professional internship experience? Can they describe a hashmap or array mutation? Can they break down a programming problem granularly enough that they can see a reasonable solution? The more investment we need to make to help them perform well in the scope of their (potential) responsibilities, the less likely it is they’ll land the role.

  • Takeaway #4: “When each link is managed separately, the system can get stuck in a ‘low-effectiveness state’…because of quality matching.”

  • Takeaway #5: “When a system is ‘stuck’, improving one link in the chain may have intended side-effect of making the overall system worse. i.e. - higher costs with the same low overall performance.”

The paradox of being a link of a stuck state system. Back to our product management scenario, if sales and engineering and marketing double their investment, we’ll have higher headcount costs, but the same product performance. It needs to be un-stuck first, which seems incredibly difficult.

  • Takeaway #6: “It’s possible to make incremental change to a ‘stuck’ chain-link system with 1) incremental improvements, 2) making change itself the objective (vs. old measures), and 3) managing the tension between decentralized autonomous action & centralized direction and coordination.”

“It takes leadership and willingness to absorb short-term losses in the quest for future gains”—juicy quote, Rumelt

Let’s sketch a plan to un-stick our made-up product management problem based on the elements of Takeaway #6 above. Let’s say the diagnosis is that product management is not delivering insight into what’s truly valuable to customers, and the dependent orgs within the company are experiencing a vacuum of leadership and product focus.

  1. Re-evaluate product leadership

  2. Suspend agile pointing + velocity measures

  3. Create the tip of an insight pipeline. Re-orient product’s output around “one-pagers”, which must articulate an insightful “job-to-be-done”, and customer interviews. Measure the production of these weekly.

  4. Develop the middle of the insight pipeline, where insight gets turned into delightful, hard-to-copy, margin-increasing features.

  5. Develop the end of the insight pipeline, which packages features into sales and marketing models

Not bad, I think. What would your plan look like in that scenario?

  • Takeaway #7: “The excellence achieved by a well-managed chain-link system is difficult to replicate”

The follows with Rumelt’s theme of cohesiveness as an impact-multiplier. If all of your strategic elements depend on and magnify one another, success will follow! If you can use your imagination, he uses IKEA as an example of excellence (highly-integrated, novel, hard-to-replicate-just-one-thing), and 2007-era GM as an example of a “stuck” system (quality-matching issues like a great drivetrain, but knobs falling off the dashboard…).

  • Takeaway #8: “…both excellence and being stuck are conditions of chain-link logic”

Last, but not least, he reinforces that a great leader must possess both the insight and fortitude to make the changes necessary to create excellence in chain-link systems. The results will not be immediate, but falling short of long-term success is worse.

🔖 Bookmark! Next time a new chapter: Using Design 🎨


The book again is Good Strategy / Bad Strategy: The Difference and Why it Matters by Richard Rumelt, and Thank you for reading Takeaways!

If you’ve hung in there for a while, thank you! I very much appreciate that you read along and I hope you find some inspiration along the way. If that’s been the case, you can help me out by sharing Takeaways with someone you think might be interested.

Here are some sharing ideas:

  1. Share a link to the summary of Never Split the Difference (which covers 100% of the book), to a friend or colleague on Slack.

  2. Mumble the words “strategy” and “proximate objective” to yourself whilst riding on public transportation.

  3. Recall one thing you’ve learned or been inspired by, and use the button below to generate some kind of social share thingy.

Share Takeaways