Interpreting a translation-cost memo
An internal memo on translation performance states: “Outsourcing rose because in-house productivity fell.” The figures are: 2023 total output 12.4 million words, of which 8.1 million external and 4.3 million in-house. 2024 total output 13.6 million words, of which 9.9 million external and 3.7 million in-house. The external rate rose from €0.11 to €0.12 per word; the internal charge-out rate stayed at €0.18 per word. External quality: 2.1 errors per 1,000 words and 6% rework; in-house: 1.3 errors per 1,000 words and 2% rework. Overtime rose 14%. The unit’s headcount was 42 at the end of 2023 and 40 at the end of 2024. The memo gives no data on hours worked, absence, or CAT tool leverage.
1. Test the memo’s causal claim. What do the figures actually support?
Read the worked answer
The figures support a shift towards outsourcing, not the memo’s causal explanation. In-house output fell from 4.3 million to 3.7 million words, a decline of 0.6 million, or about 14.0%. External output rose from 8.1 million to 9.9 million, an increase of 1.8 million, or about 22.2%. Total output rose only about 9.7%. That is consistent with more work being sent outside, but it does not show that in-house productivity fell. Output can fall because headcount fell from 42 to 40, because absence rose, because more time was spent on review or coordination, or because work was deliberately reallocated. The memo gives no hours-worked or FTE data, so the productivity claim is unsupported. The quality figures also do not establish causation: external error and rework rates are higher, but that is a quality outcome, not a productivity measure.
2. Calculate the change in the external share of total output and the change in external spend. Show your working.
Read the worked answer
External share in 2023: 8.1 ÷ 12.4 = 0.6532, or 65.3%. External share in 2024: 9.9 ÷ 13.6 = 0.7279, or 72.8%. The external share rose by about 7.5 percentage points. External spend in 2023: 8.1 million × €0.11 = €891,000. External spend in 2024: 9.9 million × €0.12 = €1,188,000. The increase is €297,000, or about 33.3% (297 ÷ 891). If internal charge-out cost is included for context: 2023 internal cost = 4.3 million × €0.18 = €774,000; 2024 internal cost = 3.7 million × €0.18 = €666,000. Total charged cost rises from €1.665 million to €1.854 million, about 11.4%, even though total volume rose only about 9.7%.
3. What two additional indicators would you request before accepting the productivity explanation?
Read the worked answer
First, request hours worked and FTE data for in-house translators, split between translation, revision, coordination, and non-project work. Productivity is output per unit of input; without hours or FTE, the memo cannot distinguish lower output from lower headcount or more time spent on non-translation tasks. Second, request CAT tool leverage and subject-mix data - for example, match rates, repetition rates, and the share of legal versus press texts. A fall in in-house output could reflect a harder text mix or reduced leverage rather than lower effort. Absence and overtime data would also help: overtime rose 14%, which may indicate that remaining staff are absorbing more work, not that they are less productive. None of these indicators alone proves causation, but they are the minimum needed to test the memo’s claim.