Expanding an existing operation
When to expand capacity and when to remove a bottleneck in an existing facility
Many expansion projects buy capacity that the rest of the plant cannot use. Before adding a line, find the step that limits output today and decide whether the cause is the process, the equipment or the way it is managed.
Short answer
Map each step of the flow with its demonstrated capacity per hour from your own records, not nameplate ratings. The step with the lowest demonstrated capacity — or the longest queue in front of it — is the current bottleneck. Classify the cause: process (settings, recipe, changeovers), equipment (condition, design limit) or management (staffing, scheduling, maintenance routines, information). If the gap to demand can be closed by fixing that step, debottlenecking is usually the smaller, faster option. Expansion is justified when demand exceeds what the whole flow can deliver even after the constraint is fixed, or when site, quality or safety limits make the existing flow unworkable. Verify the next constraint before committing.
How do you find the real bottleneck with existing data?
- List every step from intake to dispatch, including utilities, cleaning and storage
- For each step, take demonstrated output per operating hour from the best stable recent month
- Compare with average output to see how much is lost to stops and slow running
- Look for where work-in-progress or trucks queue, or where operators wait
- Confirm on the floor during at least a few representative shifts
Key point: Nameplate capacity tells you what a machine was designed for, not what your flow delivers.
Is it a process, equipment or management problem?
Most real constraints have more than one cause. Rank them by how much lost output each explains in your data.
| Cause type | Typical evidence in your reports | Usual remedy class |
|---|---|---|
| Process | Rejects or slow running tied to specific products, recipes or settings | Parameter change, method study, changeover reduction |
| Equipment | Repeated failures on one asset, rising spare-parts use, running at its design limit | Repair, overhaul, replacement or added unit at that step |
| Management | Losses vary by shift or season; waiting for material, people or information | Scheduling, staffing, training, maintenance planning |
| Utilities or site | Several steps slow together; power, water, air, cooling or space limits | Utility upgrade or site works |
Key point: Do not buy equipment to solve a scheduling problem.
How do you compare a targeted fix with a larger expansion?
| Question | Debottleneck | Expansion |
|---|---|---|
| Gap to demand it can close | Up to the capacity of the next constraint | Up to the new design capacity of the whole flow |
| Typical disruption | Short planned stops at one step | Longer works, tie-ins, possibly new building |
| Main risks | Constraint moves to the next step sooner than expected | Under-used capacity if demand or other steps lag |
| Data you need | Step-level capacity and loss causes | Demand outlook, whole-flow design, site capacity |
| Procurement shape | Often one supplier or service | Usually several suppliers and interfaces |
For a broader decision that also includes building new, see Fix, expand or build new?.
A hypothetical check of the gap
Hypothetical example — invented values with stated assumptions
Assumptions: four steps; demonstrated capacity per hour from the best stable month; demand expressed in the same unit.
| Step | Demonstrated capacity per hour | Note |
|---|---|---|
| Intake and washing | 6.0 | |
| Grading | 4.2 | Current bottleneck |
| Packing | 5.0 | Next constraint |
| Palletising | 7.0 |
Capacity gained by fixing the bottleneck = next constraint − current bottleneck = 5.0 − 4.2 = 0.8 per hour
If demand needs 4.8 per hour, fixing grading may be enough. If demand needs 6.0 per hour, both grading and packing — and possibly intake — must change, which starts to look like an expansion.
What if you cannot measure step capacity?
- Run timed observations at each step over several shifts
- Use counters, weigh scales or photo logs as temporary evidence
- Ask maintenance and shift leaders to rank where they lose most time, then check those steps first
- State in the brief which values are measured and which are estimates
What should you write down before asking suppliers?
- Target output, product mix and quality at the end of the flow
- The current bottleneck, its evidence and the next expected constraint
- Which cause type you believe applies and how you will confirm it
- Allowed downtime and site limits for any works
- The KPI that will prove the change worked, measured the same way as the baseline
Key point: Suppliers should be asked to solve the stated constraint, not to sell capacity.
Free, optional tools
Use the framework on your own, or with these free tools. Nothing here commits you to anything.
Sector-specific projects
Global B2B Group covers cross-sector industrial projects. Single-sector projects are handled in depth by the group's specialist platforms.
Frequently asked questions
Is the slowest machine always the bottleneck?
Not always. A machine that runs fast but stops often, or waits for material, can limit the flow more than a slower machine that runs steadily. Use demonstrated output and queues, not rated speed.
How do we know where the bottleneck will move next?
Look at the step with the next-lowest demonstrated capacity. Ask suppliers to state their assumption about upstream and downstream capacity in their proposal.
Can management changes alone be enough?
Sometimes. If losses vary strongly by shift or season, scheduling, staffing or maintenance routines may explain much of the gap. Test them before buying equipment.
Does Global B2B Group recommend a specific solution?
No. Our team helps you structure the requirement and compare supplier-neutral proposals. The technical and investment decision remains with you.
Describe the constraint, not the machine
Use the free project planner to record the constraint, the target and the site limits. A structured RFQ then asks suppliers to propose solutions against that evidence.
Managed RFQ support applies to projects from USD 250,000. Buyer tools are free; no price, availability, financing or outcome is guaranteed.
Sources and frameworks
Next in this series: Upgrading production lines and automation in phases: total cost, downtime and risk
Overview framework: Start with the Outcome: A Practical Framework for Better Agrifood Procurement
General procurement guidance only. World Bank documents are cited as useful methodology frameworks; they are not binding on private buyers. Confirm legal, technical and regulatory requirements with qualified advisers.