What Is SIC 32? Website Costs Accounting Explained

SIC 32 answers the question every digital business faces: we spent a fortune building the website — is it an asset or an expense? The interpretation's answer applies IAS 38's internally-generated-intangible logic to websites, dividing the project into stages with different accounting. The planning stage — feasibility studies, defining hardware/software specifications, evaluating alternatives and suppliers — is research: expensed. The application and infrastructure development stage — acquiring the domain, developing the operating software, coding, installing, testing — is development: capitalised if IAS 38's six criteria are met. The graphical design stage follows the same development logic. Content development — creating, acquiring, preparing the information that populates the site — is capitalised only where it meets the criteria (advertising-style content rarely does). And the operating stage — training, updating, hosting, maintenance — is expensed as incurred.

The IAS 38 capitalisation criteria are the gatekeepers: technical feasibility of completion, intention to complete and use/sell, ability to use or sell, probable future economic benefits (the website will generate revenue or cost savings), availability of adequate resources, and reliable measurement of the costs. In practice, the 'probable future economic benefits' and 'reliable measurement' criteria do the heavy lifting — and where internal time-tracking is weak, the measurement criterion fails and the costs are expensed regardless of merit.

The interpretation's age shows — it predates the app economy, SaaS, and headless commerce — but its logic endures: distinguish research from development, capitalise only what meets the criteria, expense the rest. IFRIC has not replaced it, and auditors apply it to modern builds (mobile apps developed for internal use follow analogous reasoning) with the same stage-by-stage discipline.

The domain-name accounting is a small but telling application: acquiring a domain name is the purchase of an intangible asset (capitalised, amortised over its useful life or treated as indefinite-lived with impairment testing), while annual registration renewals are operating expenses. The distinction mirrors SIC 32's staging logic — the acquisition creates the asset, the renewals maintain it. Entities that expense a material domain acquisition misstate both the balance sheet and the amortisation profile.

In practice, the boundary judgements cluster in three places. First, planning vs development: the feasibility workshop is planning (expense); the sprint that builds the checkout flow is development (capitalise if criteria met). Entities need contemporaneous evidence of when the project crossed into development — the IAS 38 criteria are assessed at that point, and retrospective reconstruction ('we think feasibility was resolved in March') is weak. Second, content: product descriptions and images that make the e-commerce site transact are capitalisable; the brand campaign microsite's promotional copy is advertising expense. Third, agile development: where planning and building iterate together, the entity must identify when the IAS 38 criteria were first met and capitalise only subsequent costs — a judgement requiring honest sprint-by-sprint analysis, not a blanket policy.

Cost capture is the operational backbone: timesheets or equivalent for internal staff, invoices for external developers, allocated overheads where measurable. Entities that cannot reliably measure the development costs fail the criterion and expense everything — which is why the finance function must be involved from the project's start, not at year-end when capitalisation is proposed.

UAE practice sees SIC 32 in e-commerce builds, fintech platforms, and government-adjacent portals: the amounts are often material (seven-figure builds), the auditors test the stage boundaries rigorously, and the tax computation follows the accounting (capitalised costs amortised; expensed costs deducted) subject to corporate-tax adjustments. Groups should align the accounting policy with the tax position deliberately — an aggressive capitalisation that the tax authority disallows creates a deferred-tax and cash-tax mismatch.

Amortisation and impairment complete the lifecycle: capitalised website costs are amortised over the useful life — short, in practice, because web technology obsolesces fast (2-5 years is typical; longer needs justification). The useful life is reassessed as technology evolves; a platform rebuild can strand the old capitalised balance, triggering IAS 36 impairment. Entities sometimes discover that the 'asset' was really a series of expensed iterations with one capitalised core — the impairment test forces that honesty.

Subsequent expenditure — the endless updates, redesigns, and feature additions — is expensed unless it meets the development criteria anew (a major new module with its own feasibility and future benefits, reliably measured). The default for post-launch spending is expense; capitalisation is the exception requiring fresh justification. This is where entities most often err, capitalising 'phase 2' work that is really maintenance with ambition.

Cloud and SaaS arrangements add a modern overlay: where the website runs on SaaS that the entity does not control, IFRIC's SaaS guidance (configuration/customisation costs) interacts with SIC 32 — customisation that creates an asset the entity controls may be capitalised; configuration generally is not. The analysis is fact-specific and currently one of the livelier areas of IFRS practice; entities with material SaaS implementations should take a documented position.

The internal-vs-external cost distinction affects measurement reliability: external developer invoices are reliably measurable (the invoice is the evidence); internal staff time needs timesheets or equivalent allocation — and where the entity cannot reliably measure internal effort, the criterion fails and the costs are expensed. This creates a perverse incentive to outsource (clean invoices) over in-house builds (messy time-tracking), but the standard is neutral: measure reliably, however the work is done, or expense it.

Noor Retail LLC builds an e-commerce platform in 2026 (total spend AED 2.4 million). The SIC 32 analysis:

• Planning (AED 300,000): feasibility, platform selection, supplier evaluation → expensed (research).
• Development (AED 1.5 million): coding, integration, testing — IAS 38 criteria met (technically feasible, intent to complete, e-commerce revenue probable, costs tracked) → capitalised as an intangible asset.
• Content (AED 400,000): product catalogue creation enabling transactions → capitalised (meets criteria, directly enables benefits).
• Training and launch marketing (AED 200,000) → expensed (operating stage).

Result: AED 1.9 million intangible asset, amortised over 4 years (AED 475,000/year). In 2028, a platform migration strands the asset — carrying value AED 950,000, recoverable amount nil → full impairment. The auditor's 2026 file tested the planning/development boundary (sprint records evidenced the crossover point) and the measurement (timesheets + invoices). The 2028 impairment was not a failure of the original capitalisation — it was technology moving on, which the short useful life had already signalled.

Capitalising planning costs. Feasibility and supplier evaluation are research — expensing them is not conservatism, it is the standard.

No crossover evidence. Capitalisation needs the point when IAS 38 criteria were met — reconstruct it retrospectively and the auditor will challenge.

Weak cost capture. Without timesheets and invoices, reliable measurement fails — and the costs are expensed whatever their merit.

Capitalising maintenance. Post-launch updates are operating costs — 'phase 2' needs fresh development-criteria justification.

Overlong useful lives. Web technology obsolesces in years, not decades — a 10-year life for website costs invites impairment.

Ignoring the tax follow-through. Capitalised vs expensed drives the tax deduction timing — align the accounting policy with the tax computation deliberately.

IAS 38 is the parent standard; intangible assets are the asset class. Research vs development is the core distinction. Amortisation spreads the capitalised cost; impairment (IAS 36) tests it. E-commerce functionality most clearly meets the benefits criterion. And SaaS arrangements overlay modern complexity on the classic staging.