Count the technicians your contact centre trains before you send its calls away

I have been on both sides of the decision to move customer calls out of a region. From 2013 to 2016 I built a contact centre in Cebu for a Canadian outsourcer and watched telecom and utility clients shift queues there from Nova Scotia and New Brunswick. Since 2021 I have been chief human resources officer of an Atlantic Canadian operator that was asked, in 2024, to move almost half its voice calls offshore. My position is that the usual sourcing model for a telecom contact centre leaves out its largest workforce number. It prices a call. It does not price the contact centre as the place where a telecom company finds and trains its next technicians.

This is not an argument that calls should never leave a region, or that agents abroad do the work less well. The agents I hired in Cebu were capable, well trained and often more experienced than their Canadian counterparts. Keswald sends its overnight queue and its simple billing chats to an outside provider today, on my recommendation. It is an argument about what a sourcing decision should count.

Where our technicians came from

When finance modelled the 2024 proposal, the comparison was clean: our fully loaded cost per handled contact against the outsourcer's price, with transition costs and a risk margin. The saving came to C$18.6 million a year. I asked our own records a different question. Of the 318 technicians we had hired between 2018 and 2023, where had they worked before?

The answer was that 116 had started at a Keswald headset. They had spent one to three years answering billing and repair calls, applied for field roles and joined the craft unit through our training programme. Their first-year retention in the field was 94%, against 81% for technicians hired from outside. They already knew our systems, our customers' most common faults and, above all, how a repair ticket looks to the customer who raised it. Field supervisors said they could tell a former agent on a service call within a day.

In smaller places the route matters even more. In Sydney and Corner Brook our contact centres are among the larger private employers, and for many young people there they offer the first full-time job with benefits. A technician hired from that centre is likely to stay in the region, which is where we most need technicians. An external hire who has relocated from a city is more likely to leave after two winters.

None of this appears in a cost-per-contact model. It is not hidden. It simply sits in another department's records, and a sourcing team has no reason to look for it.

Pricing the ladder in a sourcing decision

We put a number on it. Replacing the internal route with external recruitment, allowing for the retention difference and the cost of hiring and training a technician from outside, came to about C$3.2 million a year at our hiring volumes. The estimate has wide error bars, and I told the executive committee so. Even a rough figure changes the conversation, because it moves the discussion from whether the centres are efficient to what the company would have to rebuild if they went.

We also looked at repair calls separately from billing. A repair call answered by an agent who can see the field crews' schedule, and who can walk across the floor to dispatch, was resolved on first contact more often than one answered elsewhere. That gap shows up in repeat calls and in truck rolls that were never needed. It is a quality argument, but it is specific to a type of call rather than a general claim about where good service happens.

On that basis we split the work. Overnight voice and simple billing chats went to an outside provider. Daytime voice and all repair and technical support stayed in our four centres. We also made the centres a better place to stay, with schedules fixed in twelve-week blocks instead of weekly bids and a paid route into technician apprenticeship for any agent with a year's service. The clerical union agreed to amend the contracting-out clause in return for a commitment that no one would lose their job because of the change.

What we sent away, and why

The result was a saving of C$10.1 million a year rather than C$18.6 million, including lower training costs as attrition fell from 41% to 26%. Headcount fell from 1,180 to 960 through attrition and voluntary moves. Seventy-four agents entered apprenticeships over the following two years, and repeat repair calls within seven days fell from 19% to 14%.

There were costs. Our owners' representatives were disappointed, and they were entitled to be: we declined more than C$8 million a year that a defensible model said was available. The overnight handover was poor for four months, because customers who called at night had to explain themselves again in the morning. And the arrangement depends on the apprenticeship route staying open. If we stopped hiring technicians, the case for keeping daytime voice at home would weaken considerably, and I would say so.

Some colleagues in other operators tell me the pipeline argument is a defence of local jobs dressed as workforce planning. I take the challenge seriously. The test is whether the numbers survive scrutiny without the regional argument attached, and in our case most of them did: the retention gap, the repeat-call gap and the replacement cost would have mattered in a city too. The regional employment effect is real, but I presented it to our board as context rather than as the reason.

I would put two questions to a board considering a contact-centre sourcing proposal at a telecom company. First, how many of the company's technicians, network staff and supervisors began in the contact centre, and what would it cost to find them another way? Second, which calls depend on being close to the people who repair the network, and which do not? The answers will differ from one company to the next. A sourcing decision made without asking them is a decision about the field workforce, taken by people who were only looking at the phones.