Lexicon Entry

Curation vs aggregation

Curation vs aggregation. The structural distinction between editorial selection (curation) and algorithmic or comprehensive collection (aggregation). Curation excludes; aggregation includes. Both are useful; conflating them is not.

On the distinction

Curation and aggregation are different editorial practices, and the conflation of them in current consumer-internet vocabulary is a category error worth addressing carefully.

Aggregation is the comprehensive collection of items, typically organized for browse-and-search. The Apple App Store is an aggregation; the Wikipedia category page is an aggregation; the Amazon search results page is an aggregation. The goal of an aggregator is completeness: every item in the relevant universe should be present, and the editorial work of the aggregator is in the indexing and the search-and-filter affordances rather than in the inclusion-and-exclusion choices.

Curation is the editorial selection of items, with intention, from a larger universe. The goal of a curator is the opposite of completeness: the items that make the cut earn the recommendation, and the items that do not are excluded with editorial deliberation. The Top Picks Report list of five calorie-tracking apps is a curation; the App Store’s full list of calorie-tracking apps is an aggregation. They serve different reader needs and they are not interchangeable.

On the structural difference

The structural difference between the two is whether exclusion is a feature or a bug. In aggregation, the inability to find a particular item is a failure of the aggregator. In curation, the absence of a particular item is the editorial argument: the curator is asserting that the absent item is not worth the reader’s time, and the reader who trusts the curator’s judgment is taking the absence as a recommendation.

This is the structural reason that “curated by an algorithm” is, strictly speaking, an oxymoron. An algorithm can rank, filter, and personalize; it cannot, in the editorial sense, exclude. An algorithm has no editorial standpoint from which to decide that thirty of forty things in a category are not worth the reader’s attention. A person can make that decision; an algorithm can only weigh.

Why both matter

We do not mean to suggest that aggregation is a lesser form of editorial work. The major aggregators of consumer information — Wikipedia, the App Store, the major academic indexing services — perform a function that curation cannot perform, and the world is better for them. Top Picks Report is a curated publication; we are also frequent users of aggregated reference material, and most of our research process begins in aggregation and ends in curation.

The error to avoid is conflating the two. A reader who arrives at a publication that calls itself “curated” is right to expect that the publication has applied editorial judgment, defended its choices, and excluded most of the available alternatives with intention. A reader who arrives at a publication that is in fact aggregating with light editorial dressing is right to feel the bait-and-switch.