usage-pattern-interpreter

name	usage-pattern-interpreter
description	Takes raw usage data -- even a spreadsheet export or pasted metrics -- and identifies patterns, risks, and opportunities. Translates product analytics into account intelligence a CSM can act on. Use when asked to interpret usage data, analyse product metrics, make sense of a usage report, identify trends in customer behaviour, flag usage-based risks, or when a CSM has data but does not know what it means for the account. Also triggers for questions about usage analysis, product analytics interpretation, behavioural pattern detection, usage-based risk identification, or turning raw metrics into actionable insight.
license	MIT
metadata	{"author":"Stephen Rogan","version":"1.0.0","standalone":true}

Usage Pattern Interpreter

Translates raw usage data into account intelligence. Most CSMs have access to product analytics dashboards but struggle to move from "what is the data showing" to "what should I do about it." This skill bridges that gap.

How to Use

Provide whatever usage data you have:

• Dashboard exports (active users, sessions, feature usage, workflow completions)
• Spreadsheet data (CSV or pasted tables)
• Screenshots of analytics with numbers you can describe
• Specific metrics you are trying to understand ("DAU dropped 30% this month -- what does it mean?")

Also provide:

• Account context (segment, product tier, tenure, number of licensed seats)
• Any known context for the data (did they just onboard a new team? Did they lose a key user? Is this a seasonal business?)
• What you are worried about or trying to understand

Pattern Recognition Framework

Step 1: Classify the Data

Organise whatever data you have into these categories:

Category	Metrics	What They Measure
Volume	DAU, WAU, MAU, sessions, API calls	How much the product is being used overall
Breadth	Active users / total seats, features used / features available	How widely adoption has spread across the team and the product
Depth	Sessions per user, actions per session, workflow completions	How deeply each user engages when they use the product
Engagement	Login frequency, return rate, time in product	How habitually the product is part of their workflow
Trend	Period-over-period changes in any of the above	Direction of travel -- more important than any single number

Step 2: Identify Patterns

Pattern	How to Detect	What It Usually Means	Risk or Opportunity
Steady growth	Volume and breadth increasing 5-15% month-over-month	Healthy adoption. The product is spreading organically	Opportunity: expansion (seats, tier). The team is outgrowing the current configuration
Plateau	Metrics stable for 3+ months after initial growth	Adoption has reached its natural ceiling for the current setup	Risk if the ceiling is low (they are not using most of what they bought). Opportunity if the ceiling is high (time to deepen or expand)
Gradual decline	Volume or breadth decreasing 5-10% month-over-month for 2+ months	Something is changing. Could be seasonal, could be a team change, could be a competitive evaluation	Risk: investigate immediately. Gradual declines that go unaddressed become steep declines
Cliff drop	Volume or engagement drops >30% in a single period	Something happened. A key user left, a competing tool was adopted, an internal priority shifted, or there was a product issue	Risk: high urgency. Contact the customer within 48 hours. This is a potential churn trigger
Feature abandonment	A feature that was actively used has zero usage for 30+ days	The team found a better way, the feature broke, the person who used it left, or the use case disappeared	Risk if the feature was core to their value proposition. Investigate root cause
Broadening without deepening	More users logging in but each user does less	Adoption is spreading but without training or enablement. New users are exploring but not embedding the product in their workflow	Opportunity: targeted enablement for new users. Depth follows breadth if supported
Deepening without broadening	Fewer users but each user does more	Power users are getting more value but the broader team is not adopting	Risk: concentration. If 3 power users do everything and 40 licensed users do nothing, the account is vulnerable to any of those 3 leaving
Seasonal pattern	Usage dips and peaks that repeat annually (e.g., lower in December, higher in Q1 and Q3)	Normal business cycle. Not a risk signal unless the dip is deeper than prior years	No action needed -- but document the pattern so you do not misinterpret the next dip
Spike followed by return to baseline	Sudden increase (new feature launch, new team onboarding) that fades within 2-4 weeks

Step 3: Contextualise

Raw patterns become actionable when matched to account context:

Pattern	Account Context	Interpretation
Usage declining 15%	Customer's fiscal year just ended	Likely seasonal. Monitor but do not alarm. Compare to same period last year if data exists
Usage declining 15%	Champion left 3 weeks ago	Likely correlated to the departure. The champion may have been driving adoption personally. Urgent
Usage declining 15%	Customer mentioned budget review last month	Potentially deliberate reduction. May be an early signal of downsell or non-renewal
New user spike	Customer just hired 10 people	Expected. Provide onboarding resources for the new users
New user spike	No known headcount change	Investigate. Organic spread is a positive signal. But could also be a trial by a team evaluating alternatives

Step 4: Generate Insights and Recommendations

For each significant pattern detected:

## Usage Pattern Interpretation: [Account Name]
**Data period:** [date range]
**Prepared by:** [CSM name]

### Headline
[One sentence: the most important thing the data is telling you]

### Patterns Detected
| Pattern | Metric | Detail | Confidence | Urgency |
|---------|--------|--------|-----------|---------|
| [pattern] | [metric and values] | [what you observed] | [H/M/L] | [Immediate/This week/This month/Monitor] |

### Context Assessment
[How the account context (seasonal, personnel, strategic) affects the interpretation]

### Recommended Actions
1. [Action tied to a specific pattern]
2. [Action tied to a specific pattern]

### Questions for the Customer
[Specific questions to ask that would confirm or refute your interpretation]

### Watch Signals
[What to monitor in the next 2-4 weeks to see if the pattern continues, reverses, or accelerates]

Common Misinterpretations to Avoid

What the Data Shows	Common Misinterpretation	Better Interpretation
DAU/MAU ratio declining	"They are losing interest"	Could be that MAU grew (new users logged in once) while DAU stayed stable. The daily habit is fine; the new users need activation
Feature usage is zero	"They do not need that feature"	Could be they do not know it exists, cannot figure it out, or the feature requires configuration they have not done
Session duration declining	"They are less engaged"	Could mean they are more efficient. If they accomplish the same outcomes in less time, shorter sessions are a positive signal
API call volume declining	"They are using the product less"	Could be they optimised their integration to make fewer, more efficient calls. Check if outcomes are stable
Login frequency dropping but workflow completions stable	"They are disengaging"	They may be completing workflows via API or automation instead of the UI. The product is more embedded, not less

Quality Gates

• Are you interpreting patterns, not individual data points? A single day's DAU drop is noise. A 3-week declining trend is a signal. The distinction is critical
• Have you considered at least two explanations for each pattern? The first explanation is not always correct. Usage declining could be seasonal, personnel-related, competitive, or product-related. List the possibilities before concluding
• Are your recommendations specific to the pattern? "Schedule a check-in" is not a recommendation for a cliff drop. "Call the customer today to ask about the 40% usage decline and whether it correlates with the champion's departure last month" is
• Have you accounted for known context? A usage decline during December at a retail customer is expected. The same decline in March is not

Principles

• Data tells you what happened. It does not tell you why. Your job as a CSM is to find the why. The data narrows the search space; the conversation with the customer provides the answer
• Trends matter more than snapshots. A health score of 65 is concerning. A health score of 65 that was 78 three months ago is alarming. A health score of 65 that was 52 three months ago is encouraging. The direction is more important than the position
• Not every change is a signal. Usage data is noisy. Holidays, vacations, product releases, and random variation all create fluctuations that look like trends but are not. Apply the "2+ consecutive periods" rule before classifying a change as a trend
• The most dangerous data pattern is the absence of data. An account with no usage data is not healthy by default -- it is invisible. The accounts you cannot see are the accounts you cannot save