Pattern Recognition

Thinking Tools3.2 Finding the Past FactsAlso known as Expert Pattern Templates

From the chunking work of William G. Chase and Herbert A. Simon (Perception in Chess, 1973) and the recognition-primed decision model of Gary A. Klein, Roberta Calderwood, and Anne Clinton-Cirocco (Rapid Decision Making on the Fire Ground, 1986), drawn out by the interview protocol of Klein, Calderwood, and Donald MacGregor (Critical Decision Method for Eliciting Knowledge, 1989).

Turns experts' instant reads of a situation into explicit templates, each with its cues, its label, the dynamics to expect, and the move that works, so others can learn them.

Eight signs on a quiet account, checked against the silent champion template

A client has gone quiet three months before renewal. Each tag is something the account manager has noticed. Select one, then the column for what it means against the template: a cue that predicts the outcome, a look-alike that predicts nothing, or a sign that this is some other situation. The first is placed as an example. Illustrative.

Items (select one, then a box)
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern
  • Not placedPlacedFires the templatePredicts nothingNot this pattern

The quiet account

Signs
Fires the template
Predicts nothing
Not this pattern
A cue that fires itA look-alikeAn anomaly

Against the silent champion template, the sign is

Champion moved roles. The client's sponsor of the purchase moved to another division last month.Not placed yet.Champion moved roles. The client's sponsor of the purchase moved to another division last month.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.The core cue: the person who bought the product no longer owns the budget.Champion moved roles. The client's sponsor of the purchase moved to another division last month.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.Champion moved roles. The client's sponsor of the purchase moved to another division last month.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.Team logins down 40%. Daily logins by the champion's former team have fallen by forty percent in six weeks.Not placed yet.Team logins down 40%. Daily logins by the champion's former team have fallen by forty percent in six weeks.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.A cue: the people the product was bought for are drifting away from it.Team logins down 40%. Daily logins by the champion's former team have fallen by forty percent in six weeks.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.Team logins down 40%. Daily logins by the champion's former team have fallen by forty percent in six weeks.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.New CFO lists all costs. A newly hired finance chief has asked for a list of every software contract.Not placed yet.New CFO lists all costs. A newly hired finance chief has asked for a list of every software contract.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.A cue: a cost review is coming, and nobody in the room will speak for the product.New CFO lists all costs. A newly hired finance chief has asked for a list of every software contract.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.New CFO lists all costs. A newly hired finance chief has asked for a list of every software contract.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.Same industry as a loss. The client is a regional insurer, like the account lost last spring.Not placed yet.Same industry as a loss. The client is a regional insurer, like the account lost last spring.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.Same industry as a loss. The client is a regional insurer, like the account lost last spring.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.A look-alike: insurers renewed at the same rate as everyone else.Same industry as a loss. The client is a regional insurer, like the account lost last spring.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.Big logo, like the loss. It is one of the ten largest accounts, as was the last one to leave.Not placed yet.Big logo, like the loss. It is one of the ten largest accounts, as was the last one to leave.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.Big logo, like the loss. It is one of the ten largest accounts, as was the last one to leave.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.A look-alike: size did not separate the accounts that left from those that stayed.Big logo, like the loss. It is one of the ten largest accounts, as was the last one to leave.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.Shorter emails lately. Replies from the client have become brief and slower.Not placed yet.Shorter emails lately. Replies from the client have become brief and slower.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.Shorter emails lately. Replies from the client have become brief and slower.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.A look-alike: it shows up before plenty of renewals, often when the client is simply busy.Shorter emails lately. Replies from the client have become brief and slower.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.Renewal forms sent. Procurement sent the renewal paperwork for signature last week.Not placed yet.Renewal forms sent. Procurement sent the renewal paperwork for signature last week.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.Renewal forms sent. Procurement sent the renewal paperwork for signature last week.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.Renewal forms sent. Procurement sent the renewal paperwork for signature last week.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.An anomaly: the renewal is moving; the quiet is routine.Two new teams using it. Two departments that never used the product have started this quarter.Not placed yet.Two new teams using it. Two departments that never used the product have started this quarter.Fires the template: A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.Two new teams using it. Two departments that never used the product have started this quarter.Predicts nothing: It resembles past losses on the surface, but it was just as common in accounts that renewed.Two new teams using it. Two departments that never used the product have started this quarter.Not this pattern: It says the quiet has another cause. Set the template aside and read the account fresh.An anomaly: the product is spreading, and the old champion may matter less than she did.

Your sort matched 0 of 8.Your sort matched 1 of 8.Your sort matched 2 of 8.Your sort matched 3 of 8.Your sort matched 4 of 8.Your sort matched 5 of 8.Your sort matched 6 of 8.Your sort matched 7 of 8.Your sort matched 8 of 8.

Fires the template
A cue that has predicted the outcome in past cases. Expect a cost review within sixty days; find the new decision maker within two weeks.
Predicts nothing
It resembles past losses on the surface, but it was just as common in accounts that renewed.
Not this pattern
It says the quiet has another cause. Set the template aside and read the account fresh.

The sort

  • Champion moved roles: Fires the template. The core cue: the person who bought the product no longer owns the budget.
  • Team logins down 40%: Fires the template. A cue: the people the product was bought for are drifting away from it.
  • New CFO lists all costs: Fires the template. A cue: a cost review is coming, and nobody in the room will speak for the product.
  • Same industry as a loss: Predicts nothing. A look-alike: insurers renewed at the same rate as everyone else.
  • Big logo, like the loss: Predicts nothing. A look-alike: size did not separate the accounts that left from those that stayed.
  • Shorter emails lately: Predicts nothing. A look-alike: it shows up before plenty of renewals, often when the client is simply busy.
  • Renewal forms sent: Not this pattern. An anomaly: the renewal is moving; the quiet is routine.
  • Two new teams using it: Not this pattern. An anomaly: the product is spreading, and the old champion may matter less than she did.

An example

The head of account management at a software company had noticed that her two most senior account managers could tell a renewal was in danger weeks before anything showed in the data, and that newer staff could not. She asked each for times they had known at once that an account was slipping, and what they had seen first.

In their stories, trouble began when the client's champion moved to another role or left, her team's logins fell, and a new executive asked for a list of every software cost. The client going quiet was the visible part; the change of people underneath was the cue. The seniors also named what would tell them it was something else: renewal paperwork already moving through procurement, or usage growing in new teams.

She wrote it up as a one-page template, the silent champion: its cues, what to expect over the next sixty days, and the first move: find the new decision maker within two weeks. In the next two quarters, newer staff flagged nine of twelve at-risk renewals early.

Why it works

Experts recognize situations they have seen many times, and the recognition carries the next move with it. The trouble is that they store it as a picture of the situation with no rule attached, so they cannot say what they are seeing and newcomers cannot learn it by watching. Asking for specific incidents, and for the first thing noticed in each, brings out the cues the expert responds to.

Writing the template forces a distinction the expert makes without thinking: between cues that predict the outcome and features that merely resemble past cases. A client in the same industry as a lost account looks alarming and predicts nothing; a champion who has changed jobs predicts a great deal. Adding the anomaly check, what would show this is not the pattern, keeps a newcomer from applying it to every quiet client and makes the hunch testable.

The thinking behind it

William G. Chase William G. Chase, a Carnegie Mellon psychologist, showed with Herbert A. Simon in Perception in Chess (1973) that chess masters recall real game positions far better than novices but lose the advantage on random positions. The finding is evidence that expertise rests on a large store of familiar patterns, or chunks, which experts recognize at a glance. The masters' advantage disappeared exactly where there was no pattern to recognize.

Gary Klein Gary Klein, a research psychologist and pioneer of naturalistic decision making, studied fireground commanders with Roberta Calderwood and Anne Clinton-Cirocco in Rapid Decision Making on the Fire Ground (1986). Their recognition-primed decision model holds that experienced people recognize a situation as typical, take the first workable option it suggests, and test it by imagining it played out, instead of comparing options side by side. Sources of Power (1998) extends the account to other experts deciding under pressure.

Donald MacGregor Donald MacGregor, a decision researcher, set out the critical decision method with Gary Klein and Roberta Calderwood in Critical Decision Method for Eliciting Knowledge (1989). It is a structured interview that walks an expert back through a real incident in timed passes to draw out the cues, goals, and options behind each decision. The interview in this tool, which asks for real incidents and the first cue noticed, follows the same approach.

How to run it

  1. Pick a domain with clear outcomes. Choose work where wins and losses are known, such as renewals kept and lost.
  2. Interview several experts about real incidents. Ask for times they knew at once what was happening, and what they noticed first.
  3. Pull out four parts from each story. The cues, the name of the situation, what they expected next, and the move that worked.
  4. Group the stories into templates. Keep the cues and moves that recur across experts, and name each pattern.
  5. Separate predictive cues from look-alikes. Ask what a less experienced person would notice and get wrong.
  6. Add the anomaly check. Write what would tell you this is not the pattern, and what to do then.
  7. Test the templates on new cases. Track how often each fires correctly, and revise or drop the ones that misfire.

Where it goes wrong

The template records surface features. Industry, company size, or the tone of emails go in as cues because they were present, and the template fires on cases that look similar and are not. Ask of every cue whether it appeared in the cases that went well too.

The set of templates has no anomaly checks. Newcomers apply each pattern wherever it seems to fit and act with confidence on the wrong read. Make the what-would-tell-me-otherwise line mandatory.

Only the successes are counted. A template feels proven by every case it explains, while the cases it missed go unrecorded. Log the misfires and the misses as carefully as the hits, and retire templates that do not earn their place.

When to use something else

  • Recognition-Primed Decision. When the question is how an experienced person should decide in the moment, by recognizing the situation and simulating one option.
  • Chunking. When the aim is to build one's own expertise faster, by grouping information into larger units that are easier to hold.
  • Cognitive Apprenticeship. When the know-how is best passed on by an expert working alongside the learner and making their thinking visible.

Where it sits in SOLVE eX

Go to the source

  • Perception in Chess, Cognitive Psychology 4(1), 1973
  • Rapid Decision Making on the Fire Ground, Proceedings of the Human Factors Society 30th Annual Meeting, 1986
  • Critical Decision Method for Eliciting Knowledge, IEEE Transactions on Systems, Man, and Cybernetics 19(3), 1989

Back to the catalog