Skip to main content
TurnellaBeta
WFM guideTechnology

WFM technology selection

Most WFM system selections fail not because the chosen system is bad, but because the buyer did not know what they were buying. A system with every feature on the brochure and a 30% adoption rate after go-live is not a successful selection.

Must-have vs. nice-to-have features

WFM vendors bundle features aggressively. Distinguish between features that are core to WFM function and features that are premium additions designed to increase contract value. A system that excels at core features at a lower price is almost always a better selection than a feature-rich system that does the basics less well.

Must-have features

Multi-channel demand forecasting (voice, email, chat, back-office)

Without channel-level forecasting, blended agent scheduling is based on guesswork. Volume forecasting accuracy directly determines whether the schedule is fit for purpose.

Erlang C or simulation-based scheduling engine

Headcount-ratio scheduling (e.g. 1 TL per 12 agents) does not produce accurate SL predictions. Erlang C or a simulation equivalent is the mathematical basis for WFM scheduling.

Real-time adherence (RTA) monitoring with ACD integration

Without live ACD data, the RTA function is blind. An RTA module that relies on manual updates or polling with a 5-minute lag is not usable for intraday management.

Agent self-service (schedule view, leave requests, shift swaps)

Without self-service, all schedule queries and requests route through the planning team — this is a significant administrative burden that scales badly with headcount.

Standard WFM reporting (WAPE, adherence, SL, shrinkage)

If your vendor requires custom development to produce standard WFM reports, either the system is immature or the implementation is misaligned. Both are red flags.

Multi-skill scheduling with proficiency levels

Most contact centres run multi-skilled agents. A scheduler that cannot model proficiency-weighted routing efficiency will overstaff or understaff specialist queues.

Nice-to-have features

Intraday re-optimisation

Useful for dynamic schedule adjustment as volume deviates from forecast intraday, but not essential if your intraday team manages breaks and discretionary time manually.

AI/ML demand forecasting

Advanced forecasting models add value in operations with complex non-linear seasonality. In most operations, a well-configured statistical model (Holt-Winters, ARIMA) produces similar WAPE.

Gamification and engagement modules

Nice-to-have for operations with engagement challenges, but not a core WFM requirement. Do not pay for this feature if it is bundled at cost into the licence.

Integrated quality management (QA)

WFM and QA are adjacent disciplines but separate systems. Integration is useful for a single-system view but not essential — most operations manage the two platforms separately.

Fatigue and wellbeing modelling

Available in advanced systems. Valuable in 24/7 operations managing rotating shifts and accumulated rest deficit. Overkill for standard M-F daytime operations.

Automated scheduling with rule constraints

Full automated scheduling (no planner intervention required) is aspirational for most operations. Semi-automated with planner review is a more realistic target — do not pay a premium for automation that your operation cannot actually use.

The bundling trap: When a vendor includes nice-to-have features (gamification, AI forecasting, QA integration) as part of their base licence price, ask for an unbundled price — a licence that covers only the must-have features. If the vendor cannot or will not provide an unbundled price, the nice-to-have features are subsidising the must-have ones, and you are paying for functionality you may never use.

Vendor evaluation criteria

CriterionWeightHow to assess
ACD integration quality and speedHighRequest a live demo with your actual ACD platform, not a simulated demo environment. Test the latency between agent state change in the ACD and state update in the WFM system. Over 60 seconds is unacceptable for RTA.
Scheduling engine accuracyHighProvide the vendor with 6 months of historical volume and AHT data. Ask them to run a back-calculation — what schedule would their engine have produced, and what SL would that schedule have delivered? Compare with actual results.
Implementation track record for your operation sizeHighRequest references from 3 customers with similar headcount, channel mix, and ACD platform. Ask specifically about implementation timeline accuracy, go-live issues, and first 90 days post-go-live support.
Total cost of ownership over 3 yearsHighRequest a full TCO breakdown: licence fee, implementation fee, training cost, integration development cost, annual support fee, and upgrade costs. Many vendors quote a low licence fee with high implementation and support fees.
Agent self-service mobile experienceMediumAsk agents from your operation to complete core self-service tasks (view schedule, request leave, view adherence) on the vendor's mobile app without guidance. Task completion rate and time-to-complete are the measure.
Reporting configurability without custom developmentMediumAsk to build your four most important WFM reports in the demo environment without vendor assistance. If you cannot build them, or require custom development to add a basic field, that is a configuration gap.
Vendor financial stability and roadmapMediumWFM systems are 3–5 year contracts. Request the vendor's last two years of annual report or investor update. Ask specifically about product roadmap commitment — features announced on the roadmap but not delivered within 18 months should be discounted in evaluation.

Red flags during vendor evaluation

The demo is always in a pre-built demo environment

A vendor who cannot demo in your environment with your data is telling you the integration is not ready or is more complex than they are presenting. Insist on a live environment demo or discount the demo entirely.

Implementation timeline is under 3 months for a mid-size operation

A WFM implementation for a 200-seat operation typically takes 4–6 months done well. A vendor promising 8–10 weeks is either underscoping the implementation or planning to deliver a minimal configuration that will require months of post-live adjustment.

Key functionality is on the roadmap but not yet available

Roadmap promises are not contractual. If a feature is critical to your use case, require a contractual commitment to deliver it within a defined timeframe, or exclude it from your evaluation of current capability.

References are all from operations of a different size or sector

A WFM vendor with strong references in outsourced BPO may not be a good fit for a regulated financial services operation. References should be from operations facing similar constraints: similar size, similar channel mix, similar ACD platform, similar regulatory environment.

WFM technology selection questions

What are the must-have features in a contact centre WFM system?

Six must-have features: (1) Multi-channel demand forecasting — separate forecasting for voice, email, chat, and back-office at 15 or 30-minute intervals; (2) Erlang C or simulation-based scheduling engine — not headcount ratios; (3) Real-time adherence monitoring with live ACD integration; (4) Agent self-service via mobile app — schedule view, leave requests, and shift swaps without planner involvement; (5) Standard WFM reporting (WAPE, adherence, shrinkage, SL contribution) configurable without custom development; (6) Multi-skill scheduling with proficiency levels. Everything else is nice-to-have.

Related guides