https://bandit.buzz/blog/ Practical research on customer decisions that learn from results. en-us https://bandit.buzz/blog/building-a-learning-loop-with-two-api-calls/ https://bandit.buzz/blog/building-a-learning-loop-with-two-api-calls/ Fri, 10 Jul 2026 12:00:00 GMT Wire BanditBuzz with POST /decide and POST /events. Cover idempotency, a local fallback, reward maturity, mismatched identifiers, and what to add after the base loop works. https://bandit.buzz/blog/customer-decisioning-landscape-2026/ https://bandit.buzz/blog/customer-decisioning-landscape-2026/ Thu, 02 Jul 2026 12:00:00 GMT Customer decisioning products split by what they already own. Compare setup, context, reward truth, identity, execution, control, and proof by starting point, not by a winner grid. https://bandit.buzz/blog/bandit-policies-under-drift-and-seasonality/ https://bandit.buzz/blog/bandit-policies-under-drift-and-seasonality/ Thu, 18 Jun 2026 12:00:00 GMT Full-history bandit posteriors can adapt too slowly when reward rates shift. How seasonality, wear-out, sliding windows, and discounting change the history rule, and what BanditBuzz does today. https://bandit.buzz/blog/language-models-need-a-policy-for-exploration/ https://bandit.buzz/blog/language-models-need-a-policy-for-exploration/ Thu, 04 Jun 2026 12:00:00 GMT Language models can help name actions and draft setup, but current bandit tests do not support relying on them alone to balance exploration and exploitation. https://bandit.buzz/blog/how-faraday-adds-context-to-banditbuzz/ https://bandit.buzz/blog/how-faraday-adds-context-to-banditbuzz/ Thu, 21 May 2026 12:00:00 GMT Faraday resolves a person and returns a small decision-time vector. BanditBuzz snapshots it, encodes a few fixed features, and lets replay decide whether that context helps. https://bandit.buzz/blog/checking-contextual-policy-with-doubly-robust-replay/ https://bandit.buzz/blog/checking-contextual-policy-with-doubly-robust-replay/ Thu, 07 May 2026 12:00:00 GMT How person-level cross-fitting and a doubly robust estimator turn logged randomized traffic into a guarded preview of a per-person policy, including the evidence gates, overlap checks, and limits that keep the preview honest. https://bandit.buzz/blog/decision-logs-propensities-and-replay/ https://bandit.buzz/blog/decision-logs-propensities-and-replay/ Thu, 23 Apr 2026 12:00:00 GMT What to store at choice time so a future policy can be checked fairly, including offered sets, propensities, context snapshots, and why a reward table alone is not enough. https://bandit.buzz/blog/scheduling-actions-with-delayed-feedback/ https://bandit.buzz/blog/scheduling-actions-with-delayed-feedback/ Thu, 09 Apr 2026 12:00:00 GMT Planning, release, acceptance, and reward each need their own time. Start the reward clock only after a direct destination accepts the action. https://bandit.buzz/blog/timing-estimand-before-timing-model/ https://bandit.buzz/blog/timing-estimand-before-timing-model/ Thu, 26 Mar 2026 12:00:00 GMT Send-time optimization can mean whole-policy lift, timing lift conditional on acting, or repeated act-or-wait pacing. These are different questions. BanditBuzz answers the first two with nested controls and does not yet run the third. https://bandit.buzz/blog/live-holdouts-and-confidence-sequences/ https://bandit.buzz/blog/live-holdouts-and-confidence-sequences/ Thu, 12 Mar 2026 12:00:00 GMT How BanditBuzz measures whole-policy lift with a stable person holdout, matured decision units, absolute and relative effects, and a 95% confidence sequence built for repeated dashboard viewing. https://bandit.buzz/blog/delayed-rewards-identity-resolved-attribution/ https://bandit.buzz/blog/delayed-rewards-identity-resolved-attribution/ Thu, 26 Feb 2026 12:00:00 GMT How BanditBuzz joins a choice made on one identifier to a result that arrives days later on another, using pending windows, nightly identity stamps, and latest-eligible-decision attribution. https://bandit.buzz/blog/thompson-sampling-useful-first-policy/ https://bandit.buzz/blog/thompson-sampling-useful-first-policy/ Thu, 12 Feb 2026 12:00:00 GMT Why BanditBuzz’s free tier uses Beta-Bernoulli Thompson sampling: no training set, no exploration-rate knob, matured counts, Monte Carlo propensities, and the limits that come with the choice. https://bandit.buzz/blog/why-do-nothing-belongs-in-the-action-set/ https://bandit.buzz/blog/why-do-nothing-belongs-in-the-action-set/ Thu, 29 Jan 2026 12:00:00 GMT Without a no-action choice, a learner can rank messages while never learning whether contact helps. How BanditBuzz treats nothing, holdout, cooldown, frequency, and timing as different jobs. https://bandit.buzz/blog/what-next-best-action-should-mean/ https://bandit.buzz/blog/what-next-best-action-should-mean/ Thu, 15 Jan 2026 12:00:00 GMT A next-best-action system should own one bounded choice among approved actions, including no action, tied to one business result. Strategy, content, consent, and execution stay with you.