Combining two sourced facts into a third claim is the fastest way to write something no source actually said. Wikipedia's policy against original research has a name for the specific failure — WP:SYNTH — and it slips past more reviewers than outright fabrication does, because every sentence in the paragraph is individually true.
The pattern looks harmless on the page. A profile cites a local paper reporting that a manufacturer opened a second plant in 2019, then cites an industry report noting the sector grew forty percent that year, and lets the two sentences sit next to each other so a reader draws the connection. Neither source draws it. The industry report never mentions the manufacturer; the local paper never mentions the sector's growth rate. The link between them belongs to whoever wrote the paragraph, not to a citation.
Reviewers who catch this don't read a paragraph as a unit — they read it clause by clause, checking each one against the footnote attached to it. A sentence built from two well-sourced halves still fails if the seam between them, the word "which," "following," or "as a result," has no source of its own. The usual fix is a {{synthesis}} tag or a flat removal, and the paragraph rarely comes back looking the same. Reinstating it clause by clause, each half attributed only to what its own source says, almost always survives; reinstating the connective tissue does not, unless a new source is found that states the connection outright.
This is easy to confuse with something that's actually fine: citing three separate profiles that each independently state a person founded a company. Three sources restating the identical claim add up to repetition, which is exactly what notability guidelines ask for. Synthesis starts only once the cited facts differ from each other and the sentence built from them says something none of them does on its own. Two sources describing two different things support two separate claims, written as two separate sentences rather than stitched into one.
Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources.
The practical test travels well beyond funding rounds and sector reports. Any sentence with a hidden "therefore" — an award followed by a sales figure, a hire followed by a stock move, a conference appearance followed by a partnership — deserves the same question before it goes in a draft: does one cited source state this connection, or does the connection only sound plausible once the two facts sit next to each other? If the answer is the second one, a better source won't fix it — only splitting the sentence will, so each half carries just what its own footnote can hold.
