Planning the professional services contact centre — low volume, high value, when Erlang breaks

Operations · Sector guide · ~7 minute read

Professional services and legal contact handling inverts most contact-centre economics. Volume is low; value per contact is high — a single new enquiry may be worth thousands in fees, and a single mishandled call may lose a client worth far more. The standard planning toolkit assumes scale that simply is not there, and the planner who applies it anyway produces confident nonsense.

Low volume, high value — different economics entirely

In a large operation, the unit economics run on cost per contact; in a firm taking thirty calls a day, that metric is close to meaningless. What matters is capture and conversion: the new-business enquiry that rings off after ninety seconds was not a 0.4% abandonment statistic, it was a conveyancing instruction or a litigation matter walking to the next firm on the search results. The value asymmetry is extreme — the marginal cost of answering well is trivial against the fee value of a single captured matter.

The planning objective therefore shifts from efficiency to availability. Occupancy targets borrowed from high-volume operations are actively harmful here: a reception or client-care team running at 85% occupancy has no headroom for the lumpy arrivals that low volume guarantees, and the contacts it drops are the expensive ones. The disciplined posture is to staff for capture — accept apparently low occupancy as the price of never missing the call that pays for the month.

When Erlang breaks — the small-numbers problem

Erlang models assume arrival volumes large enough for averages to mean something. At twenty or forty calls a day they do not: the average interval conceals a lumpy reality of three simultaneous calls followed by an hour of silence, and an Erlang calculation will cheerfully recommend 1.3 agents for a queue where the real choice is between two people and a missed matter. This is the small-numbers problem, and it is the same mathematics that makes pooling powerful in reverse — small groups are disproportionately fragile, and splitting them makes everything worse.

Three practical consequences. First, plan with coverage rules and scenarios rather than raw Erlang outputs — who answers if two calls land at once, and who is the named overflow. Second, pool wherever professionally possible: one cross-trained client-care team across departments beats four departmental part-timers on every measure. Third, respect the power-of-one effect at this scale — in a four-seat team, one absence removes a quarter of the capacity, so the absence plan is not an appendix, it is the plan.

Matter-driven demand — the case file is the forecast

Professional services demand does not follow marketing campaigns or weather; it follows the matter lifecycle. Conveyancing contact clusters around exchange and completion days — Fridays, month-ends, and every stamp-duty or incentive deadline the government invents; litigation contact spikes around hearing dates and disclosure deadlines; probate generates waves at predictable stages of administration. Each milestone in the case-management system is a future contact event with a date attached.

That makes the practice-management system the planner’s forecast engine. Completions booked for the last Friday of the month are known weeks ahead; court listings are diarised; quarter-end and tax-deadline clusters repeat annually. A simple driver-based model — contacts per matter by stage and matter type, laid over the milestone calendar — outperforms any time-series fit at this scale, because the history is too thin to model but the pipeline is sitting in the database. The disciplined planner reads the case files, not the call logs.

The qualified-staff constraint — callbacks and appointments, not queues

The deepest structural difference: the people clients actually want to speak to are fee earners, and you cannot flex a solicitor’s diary like an agent shift. Their time is billable, diaried weeks out, and consumed by the casework that is the firm’s actual product. Staffing the phones with qualified professionals waiting for inbound calls is economically absurd; leaving clients unable to reach them is commercially fatal. The queue model does not survive this constraint.

The answer is to stop running a queue and start running a capture-and-commit model. A skilled front layer — client-care handlers or paralegals — answers everything, resolves what it can, and converts the rest into booked callbacks and appointments placed directly into fee-earner diaries with protected slots for client contact. The planning discipline then shifts to the promise: callback windows sized honestly against diary capacity, kept visibly, and measured. A kept two-hour callback promise beats a sixty-second queue answer from someone who cannot help; a broken promise is worse than either.

SLA by client, not aggregate service level

An aggregate 80/20 service level is the wrong instrument for a firm whose ten largest clients may represent half its revenue. Key clients carry bespoke expectations — named contacts, agreed response times written into engagement letters or panel agreements, escalation routes to partners — and an operation can hit its aggregate target every week while quietly failing the one institutional client whose panel review is approaching.

Plan and measure at client level: each significant client’s commitments documented as explicit SLAs, routing that recognises the client and reaches their named team, and reporting that shows performance per client rather than burying it in the average. Where panel agreements or regulatory obligations (SRA service and complaint-handling expectations among them) set response standards, treat those as floors in the capacity plan. Not legal advice — validate client commitments and professional obligations with the firm’s compliance lead.

The disciplined professional services posture

Staff for capture, not occupancy — the missed enquiry costs more than the idle minute. Replace raw Erlang outputs with coverage rules, pool the front layer across departments, and read the forecast out of the case-management system where it has been sitting all along. Run callbacks and appointments instead of queues, size the promises against real diary capacity, and plan to each significant client’s commitments rather than a flattering aggregate. Small numbers forgive nothing; the discipline is in the structure, not the spreadsheet.