In-house vs outsourced support: what should you hand over?

You want your evenings back. You also know that customer conversations are where you learn why people return the product. Outsourcing customer support raises both questions: who will do the work, and how will you keep hearing what matters?

Our view is that you can hand over a defined queue and still own product feedback, exceptions and service standards. The decision should depend on workload, coverage and knowledge transfer, not on a universal ticket-count threshold or a promised savings percentage.

Formal standards draw the same line. ISO 18295-2, the international standard for organizations that use customer contact centres, sets requirements for the brand as the client, and it applies whether the centre is in-house or run by a third party. The provider can run the queue; the brand still has to manage the arrangement so customer expectations are consistently met. (The 2017 edition is currently being revised.) If you serve EU customers, data law points the same way: under Article 28 of the GDPR, a business that has a provider handle customer data on its behalf may only use providers that give sufficient guarantees, and must set the terms in a binding contract.

The founder’s side of the tradeoff shows up in an older ecommerce discussion on Reddit. Some contributors value answering customers themselves; others describe combining an external team with internal oversight. Treat it as anecdote, not a survey, but the tension it describes is real.

What are you trying to get back?

Separate interruption from workload. A small number of calls scattered through the day can disrupt a founder even when total handling time is low. A predictable email queue might be easier to delegate or schedule. A complex product incident can require engineering regardless of who answers first.

Write down the actual problem: uncovered weekends, an expanding language mix, delayed replies, or a founder who approves every refund. Each calls for a different response. An extra agent cannot fix a missing approval rule.

Map the queue before requesting a quote. The outsourcing cost guide explains the input data and billing definitions to use.

Which work can move outside the company?

Start with a simple test: could another trained person resolve the case with the documentation and authority you provide?

Work	What must exist before handover
Order and delivery enquiries	Reliable order access and a delivery exception path
Routine return requests	Applicable policy, approval limits and warehouse contact
Product setup questions	Model-specific instructions and escalation criteria
Refund exceptions	A named brand owner with authority to decide
Suspected product faults	Evidence requirements and a route to engineering

The first three may suit an external frontline team. The last two often need internal decisions even when the external team collects information and keeps the customer updated.

Technical complexity does not automatically rule out outsourcing. It makes training and a usable tier-2 escalation path more important. Keep the work in-house if you cannot provide those conditions, or if transferring the knowledge would cost more than it returns.

Compare full costs without counting the same expense twice

For in-house delivery, list pay, employer costs, tools, recruiting, training, absence cover and supervision. Use your actual payroll and local employment costs. A generic salary multiplier cannot reliably represent every location or employment arrangement.

For outsourced delivery, list the contract fee, excluded tools, setup, ongoing training, internal oversight and escalation work. Clarify who pays replacement-agent costs. Work does not disappear simply because it moves off your payroll.

Use the same coverage schedule, language mix and expected service quality in both columns. Do not compare one internal employee’s working week with an external team’s round-the-clock availability.

Keep cash savings separate from time freed up. If a founder spends fewer hours answering tickets, that is capacity regained, not automatically lower cash spending. If a manager stays on payroll, do not claim their whole salary as saved.

How do you keep the voice of the customer?

Keep access to conversations and join the review of unresolved cases. Ask for recurring product questions, return reasons and policy exceptions with enough context to act. A list of average metrics alone will not show why a customer could not pair a device or understand a warranty.

Assign an internal owner to turn those findings into changes. The provider can identify repeated confusion; someone still has to correct the listing, manual or product.

Read routine cases as well as complaints. Positive experiences and unexpected use cases can teach the product team something too. Build this feedback loop on purpose instead of relying on informal updates.

What should a pilot prove?

Choose a defined channel, product line or coverage window. Record the baseline and agree how you will judge response usefulness, resolution, repeat contacts, policy accuracy and internal supervision time.

Write down stop or escalation conditions before launch. Examples include unauthorized refunds, missing customer updates or technical cases that repeatedly return without new evidence. Set the actual thresholds with the people accountable for delivery.

Review the difficult handoffs together. Expand only when the operating arrangement works for customers and your team, not because the pilot end date has arrived.

The outsourcing onboarding checklist turns that pilot scope into a knowledge pack, access checks, assigned decision owners and go-live evidence.

The first decision to make is small: identify one queue you could document well enough to hand over, and one responsibility you should keep. Price and test that boundary before outsourcing the whole customer relationship.