Consumer Electronics Data Scraping

Published model numbers are parsed into base model plus suffix, the variant type is classified, and regional variants are linked to a base model identity while retaining their published number and market.

That linkage is what makes cross-market comparison possible without pretending the numbers are identical. Resolution runs at about 94.7%, and where it is uncertain the record carries low confidence rather than a forced mapping — because a wrong mapping silently corrupts every market comparison it appears in.

Flagged, never merged. Exclusives exist specifically to defeat price matching, and merging them into the base model makes a retailer appear to undercut when they are selling a different SKU.

We link exclusives to the base model as a relationship, so you can include them when assessing a retailer's overall price position and exclude them when checking positioning breaches. Both are legitimate analyses, and only unmerged data supports both.

Yes, where we were collecting at launch. We record launch price and date, weeks since launch and decay percentage, producing observed decay curves by category, brand and market.

Critically, generation transitions are separated. Conflating a model with its successor produces a decay curve that is actually a generational replacement, which is a common and misleading artefact in this category. For models launched before our archive begins, we report the archive start rather than implying launch coverage.

By identifying the bundled items and pricing them against their standalone listings at the same retailer, which yields an implied bundle discount rather than treating the bundle price as a product price.

Where a bundled item has no standalone listing — sometimes deliberately — we flag that the bundle cannot be decomposed rather than estimating a value. An estimated standalone price would make the implied discount fictional.

Extracted and normalised into category-specific structured fields with units standardised, and the published text retained so every value is auditable.

Retailers do sometimes disagree about the same base model, usually through their own listing errors. We surface the disagreement rather than silently picking a winner — and that disagreement is often itself useful, since incorrect specifications on your product at a retailer is a listing quality issue worth fixing.

Where a retailer exposes it publicly, yes — typically through a store availability checker on the product page. Coverage varies considerably by retailer and market.

Store-level collection multiplies volume by store count, so we scope which retailers and which store sets justify it rather than applying it universally. For most pricing use cases online availability is sufficient; store stock matters mainly for launch and allocation analysis.

Faster than most retail categories. Competitive repricing is frequent, promotional cycles are intense, and prices around launches and major sale events can change several times a day.

Daily is the baseline; hourly is worth it around launches, Black Friday-type events and for models where you actively reprice. Weekly collection will miss most of the promotional activity entirely.

Where published, yes, and it is useful as a discount reference. But RRP is inconsistently published across markets, sometimes stale, and in some categories effectively notional.

We deliver RRP as its own field rather than computing discount depth against it as the primary metric, because a discount from a notional RRP is not a meaningful figure. Decay from observed launch price is generally the more reliable reference and we deliver both.

We quote individually. The drivers are retailer count, market count, model or category scope, refresh frequency, and whether store-level stock collection is required.

A defined category across major retailers in one or two markets at daily refresh sits at the lighter end. Multi-market coverage with hourly refresh around launches and store-level stock sits higher. One scoping call, a free pilot on your own model list within 48 hours, then a fixed monthly quote. Request a quote.