Getting Started with Process Simulation
A practical onboarding guide for Business Process Improvement analysts — built around the CoffeeShop model walkthrough.
As a BPI analyst, you already know how to map a process. The question that always comes next is: so what? You can see a bottleneck, but you can't easily prove its impact — or test whether a proposed fix will actually work — without disrupting real operations.
That's exactly what Quodsi answers. It takes the Lucidchart diagram you've already built and turns it into a live simulation model. You set the parameters — how fast customers arrive, how long each step takes, how many staff are available — and Quodsi runs thousands of virtual customers through your process in seconds.
You get data on wait times, throughput, and utilization without stopping a single real transaction.
This guide walks you through the CoffeeShop model, a simple but complete example that illustrates every core concept you'll need for real-world BPI work.
Opening Quodsi
Open your Lucidchart document. In the top-right corner of the toolbar, you'll see a small orange Quodsi icon. Click it. The Quodsi panel opens on the right side of your screen, labeled "Quodsi Model."
The panel context changes based on what you click: blank whitespace shows model-level settings; a shape shows that component's panel; a connector/arrow shows its routing panel. Clicking the orange icon again toggles the panel open or closed.
If the diagram page has never been converted, the panel will prompt you to transform it into a simulation model. We'll cover that in Section 3. First, let's look at the CoffeeShop diagram — which has already been converted — to see what a working model looks like.
The CoffeeShop Diagram
The diagram has six shapes organized into two rows: a process flow on top, and shared resources below.
This maps directly to a real coffee shop process. Customers arrive, a FrontCounterWorker takes their order, a Barista makes the drink, and the customer picks it up and leaves. The simulation will tell us: How long do customers wait? How busy are the staff? What happens to the queue if one barista calls in sick?
Understanding Conversion & Mapping
When you click "Convert Automatically," Quodsi reads your diagram and assigns each shape a component type. You can review this mapping at any time via the ⋮ menu → Diagram Mapping.
| Shape Name | Component Type | What It Represents |
|---|---|---|
| Customers Arrive | Generator | Entry point — where entities are created |
| Place Order | Activity | A timed step that uses a resource |
| Make Drink | Activity | A timed step that uses a resource |
| Pickup the Drink & Exit | Activity | A timed step (no resource required) |
| Barista | Resource | A person who performs work |
| FrontCounterWorker | Resource | A person who performs work |
| Arrows between shapes | Connectors | Paths entities flow along |
If Quodsi misclassifies a shape, open ⋮ menu → Diagram Mapping, use the dropdown to correct the type for that row, then click Apply.
Configuring the Overall Model
Click anywhere on the blank white canvas — not on any shape. The panel switches to the Model view and shows a summary such as "3 Activities · 2 Resources."
Settings Tab
| Field | What It Controls | CoffeeShop Setting |
|---|---|---|
| Model Name | Descriptive label for this simulation | "CoffeeShop — Current State" |
| Run Time | How long the simulation runs (Minutes / Hours / Days) | 24 Hours |
| Replications | Number of independent runs to average together | 1 (use 10–30 for final analysis) |
| Clock Unit | Base time unit for all durations in the model | Minutes |
| Warmup Time | Startup period excluded from result statistics | 0 Hours |
Because the simulation uses randomness, each run produces slightly different results. Running it 30 times gives you 30 data points that Quodsi averages together. For initial exploration, 1 replication is fine. For a final stakeholder analysis, use 10–30 replications.
Warmup lets the model "open for business" before statistics are collected — giving the system time to reach a realistic steady state before you start keeping score. For simple models, 0 is fine to start.
The Generator — Customers Arrive
Click on the "Customers Arrive" shape. A Generator is the entry point for your process — where entities are created and injected into the flow. Every model needs at least one.
Time Between Arrivals — The Key Setting
| Field | CoffeeShop Value | What It Means |
|---|---|---|
| Duration Type | Exponential | Random arrivals with natural variability |
| Scale | 2.5 | Average of 2.5 minutes between arrivals |
| Time Unit | Minutes | — |
The Exponential distribution is the standard choice for modeling random arrivals. It captures the reality that some customers arrive close together while others arrive with longer gaps — far more realistic than a fixed interval.
To simulate a busier period, reduce the Scale value (e.g., 2.5 → 1.5 minutes). To model a slow period, increase it. This single field lets you model current state, peak load, and future growth.
Statistical Distribution Reference
| Distribution | Parameters | Best Used When |
|---|---|---|
| Constant | Value | Duration is always exactly the same (automated steps) |
| Uniform | Min, Max | Any value in range is equally likely; clear bounds exist |
| Triangular | Min, Mode, Max | You have best-case, most-likely, worst-case estimates |
| Exponential | Scale (= mean) | Random arrivals; high variability; memoryless events |
| Normal | Mean, Std Dev | Natural variation clustered symmetrically around an average |
Activities — Where the Work Happens
Click on "Place Order." Activities are the workstations, steps, and tasks in your process — anywhere time is spent and work is done. The Activity panel has five tabs.
Settings Tab — Activity Capacity
| Activity | Capacity | Implication |
|---|---|---|
| Place Order | 1 | One customer at the counter at a time; others queue |
| Make Drink | 10 | Up to 10 drinks in preparation simultaneously; less of a bottleneck |
| Pickup & Exit | 1 | One customer collecting their order at a time |
Actions Tab — The Heart of an Activity
Actions define what happens during an activity and how long it takes. The most common action type for BPI work is Delay with Resource.
Implements the seize-delay-release cycle: grab a resource, spend time doing the work, then free the resource. A FrontCounterWorker is occupied for the duration of taking the order, then freed up for the next customer.
The Place Order action uses a Triangular distribution:
| Parameter | Value | Interpretation |
|---|---|---|
| Minimum | 0.6 min | Fastest possible — everything goes smoothly |
| Mode | 1.0 min | Most likely duration — a typical interaction |
| Maximum | 1.8 min | Longest possible — indecisive customer, complex order |
This three-point estimate will feel familiar — it's the same approach as PERT estimating. If you've done time-and-motion studies, your best-case / most-likely / worst-case observations map directly to Minimum / Mode / Maximum.
Full Action Type Reference
| Action Type | What It Does |
|---|---|
| Delay with Resource | Seize → wait → release. Most common for process steps |
| Seize Resource | Acquire a resource; entity waits if all units are busy |
| Release Resource | Free previously seized resources |
| Delay | Wait for a duration with no resource usage |
| Branch | Conditionally route entities based on state (e.g., 30% require rework) |
| Split Entity | Replace one entity with multiple for parallel processing |
| Join Entities | Wait for matching entities before proceeding |
| Assign State | Modify entity state variables during processing |
| Loop | Repeat a block of actions multiple times |
| Dispose Entity | Immediately terminate an entity |
| Create Entity | Spawn a new entity while the original continues |
Routing Tab
With a single outgoing connection, routing is automatic. For decision points (e.g., "70% of customers go to Pickup; 30% require a correction"), add a second outgoing connector and set Routing Type: Probability with relative weights.
Other Activity Tabs
Financial Tab: Enable Financial Tracking to model labor costs per step — useful when quantifying the cost impact of a proposed change for a business case.
Failure Simulation Tab: Enable to model equipment downtime, such as an espresso machine breakdown. Particularly valuable for manufacturing and equipment-intensive processes.
Resources — Staff & Equipment
Click on "Barista." A Resource represents any person, machine, or equipment that activities require. Resources have limited capacity — when all units are busy, entities must wait. That waiting creates queues, and queues are what you're here to analyze and reduce.
| Field | Barista | FrontCounterWorker |
|---|---|---|
| Resource Name | Barista | FrontCounterWorker |
| Resource Capacity | 1 | 1 |
Changing Resource Capacity is how you simulate staffing changes. Want to know what happens if you hire a second barista? Change Capacity from 1 to 2, run the simulation, and compare wait times and throughput against the baseline. This is core BPI analysis: quantifying the ROI of a staffing change before you actually make it.
Running the Simulation
Step 1: Validate
Click blank whitespace to return to the model panel. Open the ▲ Validation tab (5th tab). A healthy model shows:
"Model validation passed successfully" — 0 Errors, 0 Warnings. The model is structurally valid.
Errors (red) must be resolved before running. Warnings (yellow) are worth reviewing but won't block the run.
Step 2: Run
Click the green "Run Simulation" button at the top of the panel. When complete, a scenario card appears in the ▶ Scenarios tab showing "Done" with a green checkmark, plus run date, replications, and duration.
You can store up to 5 scenarios per model page. Once the limit is reached, delete an old scenario from the Scenarios tab to free up a slot.
Viewing Results
| Button | What You Get | Best Used For |
|---|---|---|
| XLS | Excel spreadsheet with full statistics for all components | Embedding numbers in presentations; further calculations |
| Orange chart icon | Interactive view with stats overlaid on the diagram | Quick exploration; visual stakeholder summaries |
What to Look For
Running "What If" Scenarios
After establishing a Current State baseline (Scenario 1), make a change and run again. Each scenario takes less than a minute to configure.
Scenario: Add a second barista
Opens the Resource component panel
Simulates hiring a second barista
Returns to model-level panel
Generates a new scenario alongside your baseline
Review both scenario cards — compare wait times, utilization, and throughput
Scenario: Reduce order time through process improvement
Opens the Activity panel
Reveals the Triangular distribution settings
Reflect a streamlined script, digital menu, or training improvement
Scenario: Model peak demand (double the arrival rate)
Opens the Generator panel
Doubles the arrival rate — twice as many customers per hour
See where capacity breaks down under increased load
Each scenario takes less than a minute to configure and run. The result is a data-backed answer to questions your stakeholders are asking — before you've changed a single real process or hired a single real person.
You're now equipped to convert any Lucidchart process diagram into a working simulation model, configure arrivals, activities, and resources, run scenarios, and read results. Start with the CoffeeShop model, then apply the same pattern to your own processes.