COMU1130 Chap.3 Datafication, Commodification and Selection
Datafication, Commodification and Selection
Three linked mechanisms explain how platforms reorganise social activity. Datafication translates lived actions into processable signals. Selection orders visibility through ranking and recommendation. Commodification captures value from transactions, attention or data.
This chapter teaches students to follow one action through the whole chain, distinguish observed traces from inferred categories, and show how revenue incentives can feed back into what is measured and displayed. Datafication is not the transfer of an intact social reality into a database. A platform chooses fields, timestamps and proxies, so context that does not fit the schema can disappear.
An observed click, location or rating may then support an inferred interest, routine or quality category. Selection uses such categories to decide what is encountered, in what order and under which conditions. Because abundance makes some filtering unavoidable, the critical question is which definition of relevance is operationalised and whose objective it advances. Commodification includes more than charging a fee.
Attention can be sold to advertisers, information can improve another product, reputation can organise labour access and activity can increase the value of an ecosystem. The three mechanisms form feedback. A selected recommendation changes behaviour; later behaviour becomes evidence; commercial incentives influence which response the service optimises. Accuracy must therefore be read against the target.
A model can predict the next click well while doing little for learning, wellbeing or fairness. Institutional setting changes the stakes: cultural discovery, employment access and public eligibility may use similar sorting techniques but carry different duties and appeal needs.
Mechanism-first analysis keeps a judgement testable by naming the trace, proxy, output, value relation and affected actor before declaring the outcome beneficial or harmful.
What this chapter covers
- 01
translation from activity to data
- 02
observed and inferred signals
- 03
ranking and relevance criteria
- 04
attention and data as value
- 05
feedback between outputs and behaviour
Worked example · free
Trace a reading recommendation loop
- 1Name the recorded actions.
- 1Identify the inferred preference.
- 1Explain the ranking output.
- 1Locate the revenue logic.
- 1Show the feedback effect.
Key terms
- Datafication
- The translation of activities or qualities into processable data.
- Selection
- The ordering of platform visibility through algorithms and interfaces.
- Commodification
- The conversion of activity, attention, information or objects into tradable value.
Datafication, Commodification and Selection FAQ
How is datafication different from digitisation?
Digitisation converts an existing item into digital form. Datafication renders aspects of life measurable and comparable, often creating new categories and behavioural traces that did not previously operate as formal data.
Why is selection not neutral filtering?
Selection requires a criterion of relevance and an objective to optimise. The chosen signals and target create winners, blind spots and feedback, even when the calculation is applied consistently.
Does commodification always mean charging users?
No. Value can be captured through advertising, transaction fees, subscriptions, data access or attention. A service offered without a direct price can still organise user activity around a commercial exchange.
How do I keep the three mechanisms distinct in one paragraph?
Give each mechanism its own verb and object. Datafication translates an activity into a specified trace or proxy. Selection uses that information to rank, recommend or allocate a named opportunity. Commodification links the activity or output to fees, advertising, attention, product improvement or another value relation.
Then add the feedback: the selected outcome changes later action, and commercial or institutional incentives may influence what the system measures next. If the paragraph can remove one mechanism without changing its causal chain, that concept is probably decorative. Finish with the stakeholder consequence and the criterion used for judgement.
This sequence preserves the connections among the mechanisms without collapsing them into the vague statement that a platform uses data to make money.
Assessment move
Practise a five-arrow chain for each case: activity, proxy, stored signal, selection rule and value flow. Add the affected stakeholder after every arrow. If a claim jumps from data collection to harm or benefit without naming the intervening ranking, inference or institutional action, rewrite it until the mechanism is visible. Train the distinction between observation and inference with paired cards.
Write an observable action on the front and at least two plausible categories on the back; the existence of alternatives reminds you that classification is constructed. Take one case and label where datafication ends, where selection begins and where value is captured. Then reverse the loop by asking how the output changes later activity.
For timed practice, use the verbs translates, infers, ranks, channels and monetises in separate sentences before combining them into a paragraph. Test the paragraph in a second institution while holding the mechanism stable. This reveals why a recommendation in entertainment and a ranking in welfare cannot receive the same ethical conclusion.
Keep a short error log: omitted proxy, unnamed objective, value treated only as money, or feedback ignored. Finally, draft a safeguard aimed at one link in the chain, such as correcting input data, explaining a category, changing a ranking target or limiting secondary use. A recommendation that cannot be placed on the chain is probably too generic.