Problem: The client noticed a discrepancy between the estimated word count provided by WPML for a product translation and the actual word count they calculated manually. They also observed that the translation credits charged were higher than expected based on the estimated word count. Solution: 1. We verified the word count for a completed translation job and confirmed that the count matched the credits charged. 2. We examined the XLIFF file for the job in question and found that the word count included meta data from SEO, post excerpts, and texts from media on the page, which explained the higher word count. 3. We clarified that all texts within a single translation job are counted, including duplicates, but if the same texts are included in later jobs and have already been translated, they are not charged again. 4. We explained that the estimated word count shown before translation is a rough estimate based on character count and typical word length, which may not always match the actual word count. 5. We suggested that if DeepL translation costs are higher than desired, the client could consider using Microsoft for translation, which is half the price. For more details on translation costs, you can check the WPML Automatic Translation Pricing page.
