.svg)
.svg)
.svg)
%201Q26_%20GRC%202029.png)
.svg)
.svg)
Which translation platforms report spend and quality scores by vendor?
A translation platform reports spend and quality by vendor when every agency's rates, word counts and quality reviews are captured in one system, so cost and MQM scores can be cut by the provider that did the work. In Smartling, spend by vendor comes from per-agency rate cards combined with the Word Count Report's Agency filter, and quality comes from Linguistic Quality Assurance (LQA) reports that break MQM scores and error density down by locale, project and job. Both can be exported as CSV files through Smartling's Data Access Tool into Tableau or another business intelligence (BI) tool, where finance and localization teams build the per-vendor scorecard.
Last reviewed: October 5, 2026
Why is translation spend and quality hard to report by vendor?
Translation spend and quality are hard to report by vendor because rates, volumes and quality results usually live in different places, and only some of them record which vendor did the work. Five patterns cause most of the gap:
- Rates live in contracts, not in the platform. When each language service provider's per-word rates sit in a PDF, nobody can multiply volume by rate without a spreadsheet, so spend by vendor gets calculated once a quarter instead of whenever a decision needs it.
- Each vendor scores its own quality. Vendors that run their own QA on their own scale produce numbers that cannot be compared; a "98" from one agency and a "pass" from another tell a program manager nothing about which vendor is stronger.
- Vendor identity is not always a reporting dimension. Many quality reports are organized by language, project and job rather than by provider, so a per-vendor view depends on how work was assigned and how each report is scoped.
- Estimates are not actuals. A cost estimate predicts a job before translation starts, while the vendor's invoice reflects what linguists actually did, so a spend report built only on estimates drifts from what finance pays. The mechanics of that gap are covered in why translation estimates differ from invoices.
- Executives read reports in BI tools, not in the TMS. Finance and operations leaders who compare translation vendors alongside other suppliers need the data in their own BI environment, which means an export path, a refresh schedule and enough history to show a trend.
What should a vendor spend and quality report include?
A vendor spend and quality report needs five layers, and each one has to carry the vendor as a dimension or be scoped so that it does:
- Rate basis per vendor: the agreed rate for each vendor by language pair, content type, currency and workflow step (translation, editing, review, post-editing), stored where cost calculations can use it. Without this layer every other number stays in words rather than money.
- Volume attributed to the vendor: words completed by each agency, split by workflow step, target language and fuzzy-match tier, with weighted words so translation memory discounts are already reflected. Volume multiplied by rate is the spend line.
- Quality on one scale: an MQM score, error rate and error density per 1,000 words calculated from the same schema for every vendor, so a difference in score reflects a difference in work rather than a difference in grading. The scoring mechanics are explained in how translation quality scoring works.
- Delivery against due dates: on-time delivery by job and workflow step, which turns a vendor's turnaround promise into a measured rate that can sit next to cost and quality in a renewal discussion.
- An export and refresh path: scheduled CSV or file delivery with enough history to show a trend, because a vendor scorecard is read quarter over quarter, not once.
Which Smartling reports carry vendor, cost and quality data?
Eight documented reports and exports supply the raw material for a per-vendor spend and quality view:
| Report | What it adds to a vendor view | Documented detail | fuente |
|---|---|---|---|
| Rate cards | Spend basis per agency | Per language pair, currency and content type; a rate per word, per hour or non-payable for each workflow step | Smartling Help Center, "Rate Cards for Translation Costs" |
| Word Count Report | Completed volume by vendor | Filters include Agency, Linguist, Translation Language and Workflow Step Type; Weighted Words column; up to a 1-year date range per run; CSV download | Smartling Help Center, "Word Count Report" |
| Translation Activity Report | Vendor check on a single job | CSV columns: User Name, Agency Name, Language, Word Count Type, Fuzzy Score, Word Count | Smartling Help Center, "Translation Activity Report" |
| Cost estimate | Predicted cost per job | Word counts and cost per workflow step, per language, calculated from rate cards; estimates available for 60 days | Smartling Help Center, "Get an Estimate for Translation Costs" |
| LQA Dashboard | Quality score trend | MQM score, error rate and objective errors broken down by locale, project and job; data generated every 24 hours | Smartling Help Center, "Assess Translation Quality with the LQA Dashboard" |
| LQA Error Density Report | Quality evidence for vendor agreements | Errors per 1,000 words by language, job and project; PDF or CSV download and scheduled delivery | Smartling Help Center, "Linguistic Quality Assurance Error Density" |
| Data Access Tool | Export to a BI tool | 13 CSV datasets in an AWS S3 bucket, refreshed every 24 hours, each with up to three years of data; activated by the Customer Success Manager | Smartling Help Center, "Smartling's Data Access Tool for Business Intelligence" |
| Translation Memory Leverage Report | Savings context per vendor review | Words and estimated cost savings per month by fuzzy tier and SmartMatch; filters for project and locale | Smartling Help Center, "Cost Savings Reports" |
How do you build a spend and quality scorecard by vendor in Smartling?
A per-vendor scorecard in Smartling is assembled from rate cards, word counts and LQA reports in five steps:
- Load each agency's rate card - Under Team > Agencies, add a rate card per language pair, currency and content type, with a rate for every workflow step the agency performs; Account Owners can export and re-import rates as CSV to update them in bulk. Rate cards are documented for accounts that use their own agencies and translators rather than Smartling Language Services.
- Pull completed volume by agency - Run the Word Count Report filtered by Agency for the review period, split by workflow step type and language, and use the Weighted Words column so translation memory discounts are already applied. Weighted words multiplied by that agency's rate gives the spend line, and the Translation Activity Report confirms the Agency Name on any single job.
- Score every vendor on one MQM schema - Run LQA with the same MQM-compatible schema across all vendors, as the Smartling Help Center article "LQA: Overview" recommends for consistent, comparable results. When each vendor covers distinct locales or projects, the LQA Dashboard and Error Density Report breakdowns by locale and project become vendor-and-language-pair views.
- Add delivery and leverage - Through the Data Access Tool, the Due Date Variance dataset reports on-time delivery per job and workflow step, and Fuzzy Match Leverage reports translation memory reuse by project, job, locale and workflow, which explains why a vendor's cost per source word can fall over time.
- Export and schedule - Load the Word Count Report CSV, the rate card export and the Data Access datasets into Tableau or another BI tool, then schedule the Error Density Report by email so quality evidence arrives with each vendor review.
Este enfoque se adapta a equipos que...
- Run two or more translation agencies or language service providers in one platform and renegotiate rates or SLAs on evidence rather than anecdote.
- Split member-material translation across vendors by line of business, such as Medicare Advantage, Medicaid and commercial plans, and need cost and quality for each.
- Already score translations with LQA and want those MQM results next to spend in the same review.
- Report vendor spend to finance or procurement in a BI tool alongside other supplier data.
When vendor-level reporting may not be the right priority
- Programs with a single translation vendor, where per-vendor reporting reduces to ordinary program reporting.
- Teams whose open question is reconciling one invoice line by line; that is a word-count reconciliation task covered in why translation estimates differ from invoices.
- Finance teams that need intraday spend figures, since Data Access files refresh every 24 hours and the Word Count Report is generated when it is run.
Evaluation checklist: questions to ask about vendor spend and quality reporting
Which reports carry the vendor as a filter or column?
Ask to see the actual filter list on the volume and quality reports. A vendor field on the word count report but not on the quality dashboard means quality has to be scoped by locale or project to stay attributable.
Can rates be stored per vendor, per language pair and per workflow step?
Spend by vendor is only as accurate as the rate table behind it, and a single blended rate per vendor hides the difference between translation and review costs.
Is every vendor scored on the same quality schema?
Comparable MQM scores require one schema, one severity weighting and one pass threshold across all vendors; otherwise the scorecard compares grading styles rather than translations.
Does "real-time" mean real time?
Ask for the refresh cadence of each report and export. A daily refresh is enough for a monthly vendor review but not for an intraday budget alert.
How does the data reach your BI tool, and how much history comes with it?
Confirm whether delivery is a native connector or a file export, which BI tools the documentation names, and how many years of history each dataset carries.
What goes on the executive version of the scorecard?
A workable executive view shows four numbers per vendor, trended by quarter: spend, words delivered, MQM score and on-time rate. More detail belongs in the operational report underneath.
How does Smartling report translation spend and quality by vendor?
Smartling reports spend and quality by vendor by recording each agency's rates, completed words and LQA results in one platform and exposing them through reports that can be filtered, downloaded or exported to a BI tool. Per-vendor scorecards are assembled from these reports, which keeps every number traceable to the report that produced it.
Can Smartling report translation spend by vendor?
Smartling reports spend by vendor by combining rate cards with word counts. Rate cards store each agency's per-word or per-hour rates by language pair, content type, currency and workflow step, and feed the cost estimates shown on jobs. The Word Count Report filters completed work by Agency, Linguist, language and workflow step type for up to a one-year range and includes weighted words, so an agency's weighted words multiplied by its rate card gives spend for that vendor. The Translation Activity Report adds an Agency Name column for any single job.
Can Smartling report quality scores by vendor and language pair?
Smartling's LQA Dashboard reports MQM score, error rate and objective errors by locale, project and job, and the LQA Error Density Report shows errors per 1,000 words by language, job and project, which the Help Center article "Linguistic Quality Assurance Error Density" describes as input for revising vendor agreements such as SLAs. Both reports are organized by locale, project and job, so a vendor-and-language-pair view comes from scoping them to the locales and projects each vendor covers. LQA evaluations can be performed by human linguists, by LQA Agent, or both, on human, machine or post-edited translation.
Can Smartling export reporting data to Tableau, Looker or Power BI?
Smartling's Data Access Tool delivers reporting data as CSV files in an AWS S3 bucket for import into "your own Tableau instance or other Business Intelligence (BI) tool," in the words of the Help Center article "Smartling's Data Access Tool for Business Intelligence." Tableau is the only BI product that article names; because delivery is CSV files rather than a named connector, Looker or Power BI users load the same files through their own import. The 13 datasets include Job Workflow Word Counts, Due Date Variance, Fuzzy Match Leverage and Translator Quality Performance - 90 Days; files refresh every 24 hours, hold up to three years of data, and are activated by the account's Customer Success Manager.
For health plans that run several language service providers across lines of business, these reports put vendor activity in one place, the centralized model described in Smartling's guide Member communication translation centralization: a comprehensive guide.
¿Listo para ver a Smartling en acción?
Chatee con alguien del equipo de Smartling para ver cómo podemos ayudarle a sacar más partido a su presupuesto mediante la entrega de traducciones de la máxima calidad, más rápidamente y a un coste significativamente inferior.