Analytics dashboards are accessed from the Phrase Analytics button on the navigation panel.
-
Charts showing by provider include both linguists and vendors.
-
Language pairs are shown as a concatenation of source to target language.
-
Charts showing by project only display the first 10,000 projects while the full project list is included when downloaded.
-
Dashboards show data in UTC timezone.
-
Dashboards are refreshed twice daily:
-
12:00am and 12:00pm UTC for the EU instance
-
07:00am and 19:00pm UTC for the US instance
Each dashboard shows the time of the last data refresh completion. It is important to note that the displayed timestamp may not include data generated up to 3 hours prior. This is because transactions may not be captured in time when the data refresh process is triggered.
-
-
Data reports are available for the past 12 full months, allowing users to have a full year at any point in time.
A video review of the dashboards:
Note
Due to continuous improvements, the user interface may not be exactly the same as presented in the video.
Access rights for project manager roles to view advanced analytics charts and dashboards can be set by an administrator in user settings.
Every dashboard supports different filters to display selected data in the charts. In particular, the Job state filter takes into account the following job states:
-
.
-
.
When a job or a project is deleted, the job is marked with this state as long as it sits in the recycle bin.
-
.
A job removed by manually or automatically emptying the recycle bin.
-
.
Note
When a job is archived, its state changes to archived. When a project is archived, the state of the job does not change.
It is recommended that the date filter is not run for more than one year in time, as performance may degrade and chart may no longer preserve definition.
Every chart can be downloaded in .csv, .xlsx, .json and .png file formats. Click on the 3-dots button at the top right of the chart and select Download results:
-
To download formatted results including a comma separator in large numbers, select the desired file format from the window.
-
To download unformatted results, hold the Shift (Windows) or Option (Mac) key and select the desired file format from the window.
Due to different counting methods across different languages, word count as presented may not be the same as word counts produced by other applications.
-
Each join tag is replaced with one space.
-
Other tags are removed.
In languages using a whitespace for separating words (e.g., English):
-
Each number sequence including
+-,.is replaced with one character (using regexp expression[+-]?[0-9]+([., -]?[0-9]++)*+). -
Each sequence of whitespaces is replaced with one space.
-
Whitespaces at the beginning and the end of segment are removed.
-
Each sequence of characters different from space is counted as one word.
In Languages not using whitespace for separating words (e.g., Japanese):
-
Some punctuation marks are removed from the text (using regexp expression
[\u2000-\u206F\u2E00-\u2E7F\u3000-\u3004\u3006-\u301F\\p{P}]). -
Segment is split into sequences of characters belonging to non whitespace (NWS) Han, Hiragana, Katakana, and Thai scripts and sequences of characters not belonging to those scripts (WS).
-
Total number of words = (number of words from NWS) + (number of words from WS).
-
Number of words from WS is computed as for English.
-
Number of words for NWS is number of characters without whitespaces.
The word count takes into account all source words processed (uploaded) in TMS such as when a submission has been made, a job is created or when the continuous localization and update source have been done. Adjust the dashboard filter to include trashed or deleted jobs. With continuous jobs, calculation is based on all the source words in the last version. Data from shared content is supported.
The dashboard provides insight on processed word count:
-
Words used up throughout subscription.
-
Word consumption month on month.
-
How words are distributed between providers.
-
Which language pairs occupy the highest word volume.
-
Word volume split by metadata such as clients and domains.
Word count by project creators or job creators can be displayed through customizable charts available in advanced analytics.
If required, it is possible to view charts based on character or count instead of word count. Select the desired metric through the toggle to display relevant data in the charts.
This dashboard focuses on on-time vs overdue job delivery based on the job date created and job due date. The dashboard includes all on-time and overdue jobs which have a due date, and whose last status is not cancelled, declined or rejected by linguist. Data from shared content is supported.
If the job is not yet marked as completed, then the job due date is compared with the current date.
There are two prerequisites for data to be visualized in this dashboard:
-
Jobs require a due date.
-
With the exception for the chart, jobs need to follow the correct status order - new, emailed, accepted, and completed:
-
Date/time of job emailed is less or equal to the date/time of job accepted by provider.
-
Date/time of job accepted by provider is less than or equal to the date/time of job delivered or job completed by provider.
-
As a job can go through the same status multiple times, calculation is based on the first instance of assignment and the last instance of acceptance and completion.
Note
All job workflow steps are counted in the chart. For example, if the job has a translation step and a revision step, each of them is counted with their separate status (which may or may not be different).
The dashboard provides insight on:
-
The number of on-time and overdue jobs over time.
-
Tracking of job delivery based on metadata - i.e., clients and domains.
-
Which projects are on-track and which are not.
-
Which clients and providers tending to perform well.
-
The average time taken for jobs to be assigned, accepted, and completed split by language pair. This provides an understanding of the languages that typically take more or less time in each state.
-
The average editing time per workflow step.
If using Phrase Data, the status naming is different and correlates as:
-
Ultimate and Enterprise plans (Legacy)
Get in touch with Sales for licensing questions.
The dashboard focuses on LQA features and includes results from finished assessments presented in their own row. The assessment result may or may not include issues and the charts are based on the job creation date .
Dashboard provides insights on:
-
Average quality score across an organization and the number of errors over time.
-
Which categories have the highest severity LQA errors.
-
Which providers maintain a high quality score.
-
Number of errors by severity for each project.
-
How the quality score is performing by language pair and client over time.
The LQA error counts include repeated issues. Repeated and non-repeated issues are displayed separately in by project and by provider tables.
There is a toggle between LQA and Auto LQA insights.
Shared content
Data from shared content is supported. The filter identifies whether the organization is the creator of the LQA assessment or not.
Whether the provider is from my organisation or from a different organisation is identified in shared content within the dashboard:
-
If the user is from your organisation, the actual username and full name is displayed.
-
If the user is from a different organisation, the user ID and organisation name is displayed as in the scorecard.
This impacts the following provider fields in the Quality dashboard:
-
translation_provider_username -
translation_provider_full_name
-
Team, Professional, Business and Enterprise plans
Get in touch with Sales for licensing questions.
The dashboard shows Vendor quotation information for job payments and includes all quotes in new, approved, for approval, and declined status.
Quotes of deleted projects are excluded from the dashboard.
The filter identifies whether the organization is the creator of the quote or not.
-
is the minimum price set for all the jobs within the quote.
-
is the price incurred for each job within the quote.
-
shows the actual cost paid, split equally between all the job workflow steps, language pairs, etc.
-
Team, Professional, Business and Enterprise plans
Get in touch with Sales for licensing questions.
The charts show the actual spend of an organization calculated on a pricing metric (e.g. word count, character count, pages or hours) and multiplied by a price list.
The cost dashboard is based on:
-
Quotes for all the workflow steps in the project.
Note
Creating a new quote for the same workflow step overwrites the data for the previously created quote.
-
All hourly quotes for the project and the latest quote for the standard quotes for each job.
In shared projects (or quotes with assigned provider), the quote is visible to all users but the latest quote may be different for each user.
The filter identifies whether the organization is the creator of the quote or not.
When assigning a provider to a job, consider:
-
If a provider is assigned to the job during job creation, the price list from their profile is used to overwrite the project price list.
-
If a provider is assigned later, the cost is not re-calculated automatically.
Generate a quote for the assigned user to update the data accordingly.
-
If multiple providers are assigned to a job part, the first listed provider's data is used.
Generate a new quote to use any other user’s pricing information.
-
Team, Professional, Business and Enterprise plans
Get in touch with Sales for licensing questions.
This dashboard compares the raw and net price from a quote to calculate relevant savings.
When the minimum price is higher than the calculated price, the minimum price is used in the quote. In this case, the savings equal zero, as the net rate scheme is not relevant for the minimum price.
If a minimum charge fee exists, the word count calculation is no longer applied.
Note
In the legacy dashboard, when the minimum price is 0, savings are considered as the difference between raw and net price.
The filter identifies whether the organization is the creator of the quote or not.
The charts display how translation memories (TM), machine translation (MT), repetitions, and non-translatables (NT) are leveraged across jobs.
Leverage is computed as the difference between the word count in the specific match category before applying any net rate scheme and the word count in the same match category after applying the net rate scheme.
The dashboard is based on:
-
Data from default analysis.
-
Data from the last run analysis for any given job.
Data are generated only when a net rate scheme is added to the project or to the provider assigned to a job before the analysis is created. If the net rate scheme is changed in the project or analysis, the analysis displays up-to-date information.
The charts display no data in the following scenarios:
-
Projects and jobs without a net rate scheme.
-
Analyses with 0 leverage (e.g. for projects, language pairs, etc.).
-
Analyses with internal fuzzies.
-
Deleted analyses.
Only analyses created by a user belonging to the viewer's organization are included in the computation of the dashboard. Therefore, when a Buyer shares a job with a Vendor and the Vendor creates the analyses:
-
The buyer does not see the data related to the analyses in the dashboard.
-
Since the vendor is the creator of the analyses, they are able to see pertinent data in the dashboard.
-
Team, Professional, Business and Enterprise plans
Get in touch with Sales for licensing questions.
-
Team, Ultimate and Enterprise plans (Legacy)
Get in touch with Sales for licensing questions.
The following criteria is applied to the data:
-
Segments pre-translated by TM or MT are included.
-
A segment with multiple job parts is considered and counted as one segment. The sum of editing time is taken from segments across all job parts.
-
Segments with more than 30 seconds of editing time are not included in the editing time averages, but are included in the count of the pre-translated segments chart.
-
Segments from jobs where all job parts were at some point or are included.
-
Segments that were created in the previous 12 calendar months and within the current month are included.
Charts highlight:
-
This chart shows the average editing time of a pre-translated segment. For segments pre-translated by TM, it shows the editing time by TM match category score. For segments pre-translated by MT, the horizontal line shows the average editing time of MT segments. This should help to indicate the optimal value to set the threshold.
-
Segments that are not edited are included with 0 editing time.
-
TM match scores are grouped into categories (0-69%, 70-74%,..., 95-99%, 100-101%) in order to better present the data.
-
-
This chart shows the overall average editing time of a pre-translated segment from all sources. This could be used to check if the changes of the threshold have an impact. Adjusting the threshold value should result in a lower editing time.
-
This chart shows how many segments are going through pre-translation and the match category score returned by the TM.
Filters include
-
Date - which is the date the segment is created.
-
It is set to 6 months by default, to align with the criteria used for the Phrase TM threshold value recommendation.
-
-
Project template
These charts provide insights to help minimize linguistic effort by identifying the optimal TM threshold, and should serve as a starting point for capitalizing on automation. Setting several filter combinations may result in displaying a small subset of segments which may not contain enough data to provide conclusive insights.
If other dashboard criteria are required, a feature request is welcomed.