In-House vs Outsourced Data Annotation: Cost Comparison

This is the one annotation decision that is genuinely build-versus-buy, and it deserves to be argued honestly in both directions: there are programmes for which in-house is correct and programmes for which it is a two-year detour. The determining factors are utilisation, permanence and data sensitivity, not price per label.

Three observations on the CVAT example, in order of importance:

1. The outsourced figure is higher. Anyone selling outsourcing on direct price alone is arguing against a published example. The honest case for outsourcing is elsewhere.
2. The in-house figure excludes software. CVAT states that licence costs must be added to the $122,220, and the model assumes 35 annotators exist and are fully utilised for the 4.2-month duration.
3. It is one illustrative scenario. A 2.3-million-object single-modality project is not a multi-year, multi-language, multi-modality programme, and the economics diverge sharply as those variables enter.

The guide to large-scale AI data annotation cost sets out the rate ranges a vendor quote is built from.

An internal annotation team is a fixed cost against a variable workload. If your ML roadmap produces labelling demand in bursts, and most research-driven roadmaps do, you are paying for the troughs. Whether a vendor charges per unit, per hour or per project changes how that overhead appears on the invoice; see the guide to data annotation pricing models.

TELUS Digital has argued publicly that a budget built as a fixed hourly annotator rate multiplied by hours is substantially inaccurate because it excludes hidden costs incurred while ML engineers set up the annotation process. Appen has published a similar argument specifically about tooling: building an internal annotation tool introduces development, maintenance and opportunity costs, and the buyer should only build if it accepts those costs knowingly.

Both points are strongest where annotation is not the company's core business. The engineering hours spent building a labelling interface are hours not spent on the model, and that opportunity cost does not appear on any line item.

A more complete in-house model includes:

• Recruiting and onboarding, per annotator, including the ones who leave in month two
• Team leadership and project management headcount
• QA design, gold-set construction and ongoing calibration
• Annotation tooling: licence, integration, or build plus maintenance
• Storage, transfer and compute
• Security controls, access management and audit
• Workflow engineering as the taxonomy evolves
• Training refreshes after every guideline change
• Employee turnover, and the learning curve paid again each time
• Unused capacity during model-training and evaluation cycles

The QA line is the one most often under-costed; the explainer on gold sets, audit sampling and consensus shows what that process has to contain.

1. The annotation workload is permanent, predictable and strategically core. Not "we will always need labels" but "we will need roughly this much, every month, for years".
2. Sensitive data cannot leave a tightly controlled internal environment, and no vendor's residency or facility controls satisfy the requirement. Check first what a vendor's controlled annotation centres can offer.
3. The organisation already has annotation managers, QA specialists and suitable tooling. Most of the cost of building is building; if it is already built, the arithmetic changes.
4. Specialist employees must annotate as part of their normal roles: clinicians labelling clinical data, engineers labelling engineering data. Here the labour is not substitutable at any price.
5. The team can be kept highly utilised over a long period. Utilisation below roughly 70% erodes the direct-cost advantage that motivated the build.

• Demand is variable, seasonal or project-driven.
• The programme needs rapid scale that internal hiring cannot match.
• Multiple languages are required, particularly outside the top ten.
• Several modalities are in scope, each needing different tooling and different reviewer skills.
• Annotation is explicitly not something the organisation wants to become good at.
• The cost of being late exceeds the cost of the premium.

If several of these hold, the next question is which vendor; the head-to-head of Lifewood vs Sama vs Scale AI vs Appen compares managed-service models.

The common mature structure is:

• A small internal team owning the taxonomy, the gold set, adjudication and acceptance. This is the part that must not be outsourced, because it is where your definition of correct lives.
• External capacity for volume production, calibrated against the internal gold set.
• A specialist vendor, retained, for the narrow high-risk workstream where a controlled benchmark shows a meaningful advantage.

This keeps the strategic capability in-house and the fixed cost out. It also gives you a credible answer to the question that ends most in-house business cases: what happens when the person who understood the ontology leaves? An independent AI data validation layer, measuring vendor output against the internal gold set, keeps external capacity accountable.

Three things do the work. The quality framework is built in rather than assembled by the client: trained annotators, senior review, automated consistency checks and client feedback loops, against a 95%+ accuracy SLA, with below-threshold batches reworked at Lifewood's cost. A distributed workforce across 40+ delivery centres in 30+ countries absorbs volume changes that an internal team would carry as idle capacity or overtime. And coverage across 100+ languages and multiple modalities means capability can be redeployed across changing workloads rather than hired for each one separately.

Lifewood has operated in AI data since 2004, with 56,000+ registered contributors and 414,120 training hours delivered to the Bangladesh workforce in 2025, figures that describe the operating layer a buyer would otherwise have to build. Managed AI data services spanning annotation, multilingual collection, LLM training data and evaluation let one vendor relationship absorb workloads an internal team would re-staff for each time.