The Issues List for Innovation: Surfacing Bad News at Line Speed

The Vision/Traction Organizer ends not with a vision but with a complaints department: the issues list, a standing inventory of everything wrong, stuck, or merely suspected, worked through in the weekly meeting’s long hour by a three-step verb Wickman insists on — IDS: Identify, Discuss, Solve. Identify the root issue rather than the symptom; discuss it once, openly, without the meeting-after-the-meeting; solve it into a to-do with an owner. EOS’s practical rules are unglamorous and exact: keep the list visible, take the top three issues rather than skimming all twenty, and solve to done rather than to “revisit next week.”

It reads like the least strategic page of the organizer. For an innovation system it’s the most important one, for a reason the whole series has been circling: new bets die of information that existed but didn’t travel. The one-year plan’s kill conditions, the rocks’ off-track flags, the ten-year target’s falsifiability — every mechanism in this series depends on bad news reaching a decision-maker while it’s still cheap. And the base rate is against you: psychologists Rosen and Tesser named the MUM effect in 1970 — the robust finding that people delay, soften, distort, or delegate the delivery of unwelcome information. Your organization is running that bias at every layer, right now. The issues list is the countermeasure: a designated, legitimate, routine place where unwelcome information is supposed to appear — so that surfacing a problem is filing a form, not staging a confrontation.

The failure case: deviance, normalized

The canonical demonstration of what happens without such machinery is the Challenger. The night before the January 1986 launch, Morton Thiokol engineers — Roger Boisjoly among them — recommended against launching below 53°F, the coldest prior flight; the forecast was far below that, and the recommendation was reversed under pressure in a teleconference the Rogers Commission later reconstructed in detail. Diane Vaughan’s The Challenger Launch Decision gave the deeper pattern its name: normalization of deviance. The O-rings had shown anomalies for years; each anomaly that didn’t end in catastrophe became evidence the anomaly was acceptable; the definition of “normal” ratcheted, launch by launch, until the unacceptable was routine. Nobody suppressed a memo in the movie-villain sense — the system metabolized each warning into precedent.

Innovation portfolios run the identical ratchet at lower stakes: the pilot that misses its metric, explained; the churned design partner, rationalized; each explanation becoming the baseline for accepting the next miss — until a zombie project’s file is thick with individually-reasonable accommodations. Vaughan’s lesson for the issues list is structural: anomalies must be logged as anomalies — against the pre-committed evidence standards from the rocks post — before the explaining starts, because an anomaly discussed only in the presence of its explanation has already been normalized.

The existence proofs: cords, questions, and logs

The organizations that beat the MUM effect all built the same machine in different housings: they made raising a problem cheap, fast, and status-safe, and made sitting on one expensive.

Toyota’s andon cord is the purest hardware version: any worker on the line can pull the cord over any suspected defect; the team leader responds immediately, and if the issue isn’t resolved within the takt window, the line segment stops. The cultural payload is in the frequency — pulls reportedly happen thousands of times a day across a plant, and one widely-cited 2007 comparison had a Toyota plant logging on the order of two thousand pulls a week while a Ford plant managed two. At that frequency, pulling the cord carries no drama whatsoever — which is the entire trick. An issues list that gets three entries a quarter is a Ford cord: technically present, socially unusable. Amazon thought the mechanism portable enough to copy in software — Bryar and Carr’s Working Backwards describes the customer-service “andon cord” that lets frontline reps pull a defective product from sale, no approval chain required.

Andy Grove’s revolving door is the executive version. Mid-1985, with Intel bleeding in memory chips, Grove asked Gordon Moore the question he recorded in Only the Paranoid Survive: if the board threw them out and brought in a new CEO, what would he do? Moore: he’d get us out of memories. Grove: then why don’t we walk out the door, come back in, and do it ourselves? The question is an IDS technology — it identifies the real issue (we are staying in memories for reasons of history and emotion, not strategy) by structurally deleting the sunk costs from the discussion. Grove’s broader practice of constructive confrontation, and his insistence that bad news deserves faster handling than good, are the issues-list ethos scaled to a company whose strategic inflection points — his term — announce themselves first as uncomfortable data somebody has to be willing to say aloud. Ray Dalio’s Bridgewater runs a maximal variant: a mandatory issue log where recording mistakes is required and failing to log one is the serious offense — inverting the MUM effect’s incentives by fiat.

And the precondition for all of it is the thing Google measured. Project Aristotle, Google’s 2012–2015 study of 180+ teams, went looking for the composition of the perfect team and found composition barely mattered; the top predictor of team effectiveness was psychological safety — the shared belief that the team is safe for interpersonal risk — ahead of dependability, clarity, meaning, and impact. That’s Edmondson’s construct, confirmed at industrial scale, and it closes the loop this series opened in the core values post: the candor clause is what makes the issues list writable. A tool for surfacing bad news, installed in a team where the last messenger was punished, stays blank — and a blank issues list in an innovation portfolio is never good news; it’s the MUM effect winning.

Running IDS on an innovation portfolio

The mechanics, adapted. Log anomalies, not conclusions: the entry is “design partner #4 stopped responding after the pricing conversation,” not “pricing may need revisiting” — Vaughan’s rule; the interpretation happens in Discuss, in public, not in the filing. Work the top three to done: EOS’s depth-over-coverage rule matters doubly for innovation issues, because the comfortable ones (logistics, tooling) will crowd out the existential ones (nobody on the team still believes the core assumption) unless ranking is explicit. Give kill-shaped issues a fast lane: any entry that touches a rock’s kill condition or a bet’s evidence standard goes to the top, because those are the issues with a clock on them — every week of delay converts them, Challenger-style, into precedent. And audit the list itself quarterly: entries per week, time-to-solve, and the ratio of self-reported to observed problems. Grove’s inflection points and Vaughan’s anomalies both arrive first as a drop in reported issues from a team that has stopped believing reporting matters.

That’s the eighth section, and the organizer is complete — and its architecture is now visible end to end. The V/TO runs on two pages because innovation strategy is genuinely two problems: a vision problem — values that permit bets, a focus that prices them, a target held for a decade, a beachhead worth landing on, a three-year picture to steer by — and a traction problem: a one-year plan that funds evidence, rocks that contract for it, and an issues list that lets the evidence speak even when it’s unwelcome. Wickman’s hallucination line cuts both ways: vision without traction hallucinates a future, and traction without vision optimizes a treadmill. The two pages only work stapled together.