Sixty-seven percent. That is the share of consumers Adobe reported abandoning content that ran too long, from a 2015 survey of more than twelve thousand people across six countries, and it is the kind of number that survives into a headline long after the report’s method section, and its other findings, get forgotten.
Press release vs. full report: where the caveats live
A press release compresses a study to its single most quotable statistic, stripped of the qualifying detail that determines how far that statistic actually generalises. The full report is where the methodology section sits: how many people were surveyed, in which countries, over what period, and how the question itself was worded, all of which shape whether a headline number describes a broad population or a narrower one than the release implies. Reading the full document, not the press summary, is the only way to check that gap.
Reading a survey’s sample before its headline number
Sample size and composition matter together, not separately. Adobe’s 2015 survey drew from over 12,000 consumers across six countries, large enough to support broad claims; a smaller, single-country survey making an equally sweeping claim deserves more scrutiny of how respondents were recruited. A large sample does not rescue a poorly worded question or a self-selecting recruitment method, so the honest read checks both the number of respondents and how they were found before trusting the topline percentage.
Named frameworks belong to the consultancy that coined them
A framework like EY’s four categories of the sharing economy sorts a broad, messy trend into a small number of named buckets, useful for giving clients and journalists a shared vocabulary quickly. The category boundaries, though, are one consultancy’s own judgment call; no outside body sets them, which is why different firms often publish overlapping but non-identical frameworks for the same trend around the same time. A decades-old model can still hold up structurally even after its era-specific numbers stop applying: Robert Ducoffe’s 1996 concept of advertising value, built from informativeness, entertainment and irritation, is cited today for its structure, not its original print-ad figures.
Reading a forecast’s range, not just its midpoint
A workforce or technology forecast models a range of scenarios bounded by an optimistic and a conservative estimate, rather than producing one single number. The figure that ends up in most coverage is one point inside that range, and a report’s own methodology section states the assumptions driving how wide the spread actually is, which is usually where the real uncertainty in a forecast lives. The Academy’s chart chooser covers a related habit worth applying here too: matching the chart type to the actual claim being made, low-to-high bars rather than a single point that hides how much a forecast could move.