Blog · Forecast and pipeline · Professional services
Service demand forecasting turns expected work volumes and handling times into required hours. Work from a service export to a capacity comparison, then separate volume errors from handling-time errors.
Demand forecasting for services converts expected requests, appointments or assignments into the hours needed to deliver them. Compare those hours with usable staff capacity, then compare the assumptions with actual delivered work. A service count alone cannot show whether the team can absorb the workload.
Use the service register or booking export already maintained by operations. Keep one row per work item, including its ID, service type, request date, planned delivery period, completion date and recorded delivery hours. An additional forecast sheet should hold the forecast version, period, service type, expected volume and standard hours per item.
Choose the event being forecast. Incoming requests measure arrivals; scheduled appointments measure booked delivery; completions measure output. They answer different questions. A capacity plan can include opening unfinished work plus new work due in the period, but each item must appear once.
The broader demand forecasting process covers how teams agree assumptions and freeze a version. Here, the job is translating a supplied service forecast into hours and testing that translation.
Required delivery hours = Sum of forecast volume × Standard hours per item
Capacity load = Required delivery hours / Available delivery hours
A US consulting business plans November 2026 work across three services. Its forecast contains the volumes below; the later actual column contains completed work and the average recorded time for those completions. No unfinished items are included in this illustration. Hours are exact except capacity percentages, which are rounded to one decimal place.
| Service | Forecast items | Standard hours/item | Forecast hours | Actual items | Actual hours/item | Actual hours |
|---|---|---|---|---|---|---|
| Advisory calls | 240 | 0.75 | 180 | 260 | 0.80 | 208 |
| Implementation sessions | 80 | 2.00 | 160 | 70 | 2.20 | 154 |
| Review packs | 50 | 3.00 | 150 | 60 | 3.00 | 180 |
| Total | 490 | 542 |
Put these columns in A:G, with the services in rows 2:4. In Excel, enter this in D2 and fill down:
=B2*C2
In G2, enter and fill down:
=E2*F2
Calculate the forecast total independently:
=SUMPRODUCT(B2:B4,C2:C4)
Microsoft documents SUMPRODUCT as multiplying corresponding arrays and summing the results. That returns 490 hours, matching the sum of column D. Do not multiply a simple average handling time by a combined count: the service mix supplies the weights.
The staff roster and approved time plan give the following November capacity. These are supplied assumptions for the same delivery team and period as the work table.
| Capacity component | Calculation | Hours |
|---|---|---|
| Scheduled paid time | 6 staff × 160 hours | 960 |
| Approved leave | Subtract | 80 |
| Meetings and administration | Subtract | 120 |
| Training | Subtract | 60 |
| Available delivery time | 960 − 80 − 120 − 60 | 700 |
Forecast load is 490 / 700 = 70.0%. Actual delivered load is 542 / 700 = 77.4%, assuming those capacity assumptions held. The forecast leaves 700 − 490 = 210 hours of headroom; actual work leaves 700 − 542 = 158 hours.
Put scheduled hours in K2 and the three deductions in K3:K5. Available hours in K6 are:
=K2-SUM(K3:K5)
Forecast load in K7 is =SUM(D2:D4)/K6; actual load in K8 is =SUM(G2:G4)/K6.
These totals support a monthly staffing discussion. Check daily peaks, required skills and customer deadlines before assigning the apparent headroom to new work. Idle advisory capacity cannot automatically cover an implementation specialist's bookings.
Actual work used 542 − 490 = 52 more hours than planned. Split the difference using standard handling time for the volume change, then actual volume for the handling-time change:
| Service | Volume effect: (actual − forecast) × standard hours | Time effect: actual volume × (actual − standard hours) | Total change |
|---|---|---|---|
| Advisory calls | 15 | 13 | 28 |
| Implementation sessions | −20 | 14 | −6 |
| Review packs | 30 | 0 | 30 |
| Total hours | 25 | 27 | 52 |
Add H and I to the first table. The volume effect in H2 is:
=(E2-B2)*C2
The handling-time effect in I2 is =E2*(F2-C2). Fill both down.
The check is 490 + 25 + 27 = 542. A nonzero residual means the comparison uses inconsistent volumes, times or rows. Review the volume effect before changing handling standards; extra demand and slower delivery require different actions.
For a larger export, SUMIFS supports multiple selection criteria. On a sheet named Work, with period in A, service in B and recorded hours in C, this totals November advisory hours. Store period keys as text, such as 2026-11:
=SUMIFS(Work!C:C,Work!A:A,"2026-11",Work!B:B,"Advisory calls")
Covirage is not affiliated with Microsoft.
The operations lead owns the service definitions and handling assumptions; team managers own capacity. Finance checks that the same period and staff scope run through both files. Keep the demand forecast separate from decisions about staffing and delivery commitments.
Bring a service-volume forecast, completion export and staff-hours file: Covirage maps the columns, its tools compute and check the volume, handling-time and capacity comparisons, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account, which takes about a minute, then set up your business and bring a sample of your own data.
See Covirage for professional services for the supporting industry context.
Start with dated requests or bookings, service categories, completed volumes and recorded delivery hours. Add a dated volume forecast, handling-time assumptions and available staff hours. Keep open work separate from new requests so the same task does not enter requirements twice.
Use both. Requests describe customer demand, while hours translate that demand into a capacity requirement. Keep service types separate because a short consultation and a complex implementation session consume different amounts of labor.
Completed work is a useful historical input, but it can be limited by capacity. If requests were delayed or rejected, completions understate demand. Review arrivals, open work and completion dates together before treating delivered volumes as the expected workload.