What needs to be finished first
A specific variable of interest, for example ticket count, customer satisfaction, or engagement, and a suitable time window of weeks, months, or years are named.
Either quantitative data or the structured recollection of several people is available so the curve can be drawn.
What needs to be ready before start
Whiteboard or shared board with a coordinate system (X-axis time, Y-axis variable); marker pens; available data points as reference; sticky notes for events and assumptions.
One facilitator who moderates curves and guides the discussion; 3-8 participants with different perspectives on the variable; one scribe who documents events and assumptions.
Communicate the variable and time window in advance; share known data points; note that curves are hypotheses about behavior, not claims of truth.
45-90 min for one variable, separate iterations for multiple variables
Make the coordinate system clearly visible. Label the X-axis with time units suited to the variable. Label the Y-axis with variable and unit. Instruction: first draw an individual curve in silence on paper, then create a shared curve at the board. Mark events as stickies at the relevant time.
The one question this method answers
How did the chosen variable behave over time, which patterns such as growth, decline, oscillation, or stagnation appeared, and which systemic causes sit behind the pattern?
Marker: Minute
| Step | Duration | Action | Hint |
|---|---|---|---|
10-10 min | 10 min | Clarify the variable and time window. The facilitator introduces the diagram logic and shows typical patterns such as linear, exponential, S-curve, oscillating, or crash. Label the Y-axis with the variable and unit, and the X-axis with the time window. | If the variable is too abstract, for example 'culture', the curve becomes arbitrary. Operationalize it: what would be measured to make the culture visible? Agree on a measurable proxy. |
210-20 min | 10 min | Solo curve: each person draws their interpretation of the variable over the time window on paper. Mark events that shaped the curve as notes. | The solo phase prevents groupthink. Anyone who goes to the board right away starts to dominate the curve. Keep silence. For large groups, allow 2 extra minutes. |
320-45 min | 25 min | Shared curve: participants transfer their solo curves to the common diagram. If they diverge, discuss which data points and assumptions differ. Make either a consensus curve or multiple hypothesis curves visible. | Divergence is not a problem, it is gold. If everyone draws the same curve, either diversity is missing or the data are too clear. If there is disagreement, leave both curves visible and decide with data or tests. |
445-65 min | 20 min | Mark events and causes: add stickies to turning points, jumps, and bends. Which events correlate, and which causal hypotheses exist? Multiple hypotheses per turning point are allowed. | Correlation is not causality. Mark hypotheses explicitly as hypotheses. If only one explanation appears per turning point, there are probably blind spots. |
565-90 min | 25 min | Systemic reading: which feedback loops or structures explain the pattern. Connect to Causal Loop or System Archetypes. Derive leverage points or hypotheses for the next steps. | The method does not end with the diagram. Without a reading, BOTG remains a line drawing. Build the bridge to action: which hypothesis would we test to change the curve? |
What comes out at the end
Diagram with a shared curve or several hypothesis curves, turning points with events and assumptions, a systemic reading, and derived hypotheses for the next steps. Optional data snapshot as an appendix.
Date, variable, time window, and participants in the header. For follow-up workshops, attach the previous curve as a comparison and keep the curve progression visible across iterations.
Compact working template for Behavior Over Time Graph with context, input, output artifacts, and next step.
Visualizes how a relevant variable changes over time and which patterns may lie beneath it.
When and for what do we use this method?
Which data, observations, decisions, or materials are available?
Short notes along the runsheet.
Owner, date, and success signal.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Participants: 5 people from Product, Data, CS.
Feedback loop: onboarding completeness versus time to value. More completeness slows the first step, templates decouple that. Leverage point: segment-specific templates.
Special template for the 'Self-Serve Marketing' segment, test in 4 weeks, owner @lisa.
Recognize symptoms and steer against them
The curve for 'engagement' or 'culture' is arbitrary and cannot be tied to data.
Operationalize the variable with a concrete proxy metric and a unit. If no proxy is possible, BOTG is not the right tool and Mind Mapping or Rich Picture is a better fit.
One person goes straight to the board, draws, and everyone else adapts or stays silent.
Keep the solo phase strict. Draw on paper first, then transfer. Allow multiple hypothesis curves explicitly before searching for consensus.
The curve contains both measured points and felt impressions, while the difference is invisible.
Mark data points with another pen or marker on the curve. Show recollection-based parts as a dashed line. That keeps what is evidenced and what is hypothesis visible.
The curve is drawn, the workshop ends, and there is no reading or action.
Make the systemic reading and hypothesis phase mandatory. Connect to Causal Loop or concrete leverage points. Without leverage, BOTG remains a description.
Turning points are assigned to events without checking alternative explanations.
Name two or three explanations per turning point explicitly and mark all of them as hypotheses. Validate with data, interviews, or experiments after the workshop.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.