Why tasks don't get done

01Where the term comes from

“Activation energy” is borrowed from chemistry — the initial push a reaction needs before it can proceed on its own. The psychologist Mihaly Csikszentmihalyi, who pioneered the study of flow, applied the same idea to human behavior: every worthwhile activity demands an upfront burst of energy to begin, before it can become self-sustaining and even enjoyable.

The metaphor stuck because it matches lived experience: starting is disproportionately hard relative to continuing. The rigorous evidence for why lives under three other names.

02Task aversiveness predicts procrastination

The most direct evidence comes from Piers Steel's 2007 meta-analysis in Psychological Bulletin — the definitive review of the field, pooling 691 correlations. It found task aversiveness to be one of the strongest and most consistent predictors of procrastination: the more unpleasant or effortful a task feels to engage, the more reliably it gets put off.

This maps almost exactly onto TaskDriver's premise: tasks we'd rather not do are precisely the ones that stall. Reduce how aversive the first step feels, and more of them get finished.

03The brain treats effort as a cost

Underneath aversiveness is a deeper mechanism: people actively avoid effort because the brain codes it as disutility — a cost to be minimized.

The “law of less work” (Hull, 1943; extended to mental effort by Kahneman, 1973) holds that we choose actions that minimize exertion. Modern work formalizes this as cognitive effort discounting: the value of a reward is discounted in proportion to the effort required to earn it — so an upfront barrier literally makes the payoff feel smaller. And in demand-avoidance experiments (Kool et al., 2010), given a free choice, people reliably pick the lower-effort option even when it offers no other advantage.

04Removing friction beats adding willpower

If the barrier is the problem, the fix is to lower it — not to push harder. Kurt Lewin's force-field analysis showed that behavior is held back as much by small “restraining forces” as it is driven by motivation, and that the fastest route to action is usually removing a restrainer.

The modern nudge / sludge literature (meta-analyzed in Royal Society Open Science, 2023) bears this out: small reductions in friction on the initiation step reliably change whether people follow through. That is the entire design philosophy behind TaskDriver — do the digital legwork that makes a task aversive, so the remaining human step is small enough to just do.

05Priority asks a question you can’t answer alone

“How important is this?” can be answered from the task in front of you, in about a second. “What priority is this?” cannot — priority is a ranking, so answering it means holding the rest of your list in your head. The cost of the question grows with the size of your backlog, which is exactly backwards.

That isn’t a quirk of any one app. In ITIL, the framework that formalized the term for IT operations, priority is explicitly a derived value: a function of impact (how much it matters) and urgency (how soon it must happen). Jira Service Management ships it that way — two fields in, priority computed out. A task app with a single “Priority” field is asking you to compute that function in your head, at the moment you’re trying to write something down and move on.

The industry can’t hold the line either. Todoist’s own documentation defines P1 as “Most important, urgent” and P4 as “Least important, not urgent” — both axes, one scale. Linear offers Low, Medium, High and Urgent: three magnitudes and a deadline in the same menu. Microsoft never used the word at the protocol layer at all — its task field is literally called Importance. And the GTD-lineage apps mostly refuse the field: OmniFocus shipped for years with flags instead, its manual describing them as a way to “flag items of importance”.

06Urgency crowds out importance

Mixing the two axes isn’t just untidy — it feeds a documented bias. In Zhu, Yang and Hsee’s 2018 work in the Journal of Consumer Research, five experiments found what they named the mere urgency effect: people chose objectively lower-payoff tasks over higher-payoff ones purely because the low-payoff ones were framed as time-sensitive — with the rewards controlled for.

The obvious conclusion — that people are bad at prioritizing — is the wrong one. Bellotti and colleagues (CHI 2004) studied how people actually manage to-dos and found the opposite: contrary to popular wisdom, people are not poor at prioritizing, and have well-honed strategies for the task management problems they face. The weak link is the field, not the person using it. A single box labelled “Priority” invites urgency to leak in, and once it has, the app can no longer tell which of the two things you meant — so it can’t help you correct for the bias.

07What our own data shows

We looked at the tasks in TaskDriver’s own production database — 274 of them. Two things stood out.

Priority isn’t tracking deadlines. If people were using the field as a stand-in for urgency, high-priority tasks would consistently have the nearest due dates. They don’t: the median gap between creating a task and its due date runs 7 days for High, 30 for Medium and 10 for Low. There’s no clean ordering, which means the field is doing something the calendar isn’t.

Priority predicts whether the date gets honored. What it does track is follow-through. High-importance tasks land essentially on their due date; unmarked ones drift by nearly a week.

Completion rates are for dated, non-recurring tasks; recurring work is excluded because its due date is the next occurrence, which distorts any measure of lead time.

That’s the signature of an importance marker, not an urgency one. It doesn’t say when something will happen — it says whether the date you set will survive contact with the week.

08So we changed the question

If a task’s date already carries urgency — and in TaskDriver it does, down to the hour and the expected duration — then the priority field can only be carrying importance. Which means the ranking doesn’t need to be typed at all. It can be worked out.

That’s what Workload Optimizations does, currently in beta. The field becomes Importance, urgency is derived from the date, and the order to work in falls out of the two: one ranked list split into Now, Next and Later, with Now capped so it stays a set you can take from rather than another list to read. The two axes also give you Eisenhower quadrants for free — computed, so they stay true as the week moves instead of going stale the day after you file them.

And they surface the quadrant every system drops: important, but with no deadline. Those tasks appear in no date-driven view, which is precisely why they never get done. TaskDriver now notices them and asks the only question that actually helps — not how important is this, really?, but when?