What we review
Failed-attempt pattern study
A focused reading of “unable to deliver” events: which codes repeat, which streets and time windows they cluster on, and whether a second knock was ever recorded.
A fail code named “recipient not in” can mean a locked gated community in Bandar Puteri, a lunch-hour shop lot in Seri Kembangan, or a rider who marked the stop from the next street. The code alone does not settle it.
This study is for desks that are tired of arguing from memory. We count, then we sit with the people who closed those stops, then we write what the counts can actually support.
Who it is for
Dispatch leads who see a high share of failed attempts and cannot tell from the status list whether riders are missing the recipient, arriving outside the notice window, or closing the stop too early.
What you leave with
A short study note with a table of the most common fail codes, three to five street-level clusters, and a recommended change to how the desk asks for a reattempt.
Scope
One hub’s failed attempts across an agreed week. We do not ride along unless you specifically request a street sample as an add-on.
Included
- Intake of your fail-code list as the riders actually see it
- A counted sample of failed attempts from the agreed week
- Notes on whether a second attempt or a locker redirect was logged
- A study note in English and a one-hour desk briefing
Not included
- Retraining every rider
- Rewriting the fail-code list inside the app
- Doorstep photography
Who does the work
Lim Wei Han traces the sample from the Puchong desk, with a half-day at your hub if the codes only make sense beside the exception cage.
How the work runs
- You send a de-identified extract of failed attempts for one week.
- We group them by code, hour, and locality.
- We sit with dispatch to test three explanations against the floor’s memory of that week.
- We issue the study note and brief the supervisor.
Place and delivery
Mostly desk-based in Puchong, with an optional hub half-day in the Klang Valley.
What to prepare
A clean extract with timestamp, fail code, locality, and whether a reattempt exists. Remove recipient names and phone numbers before you send it.
Constraints
If the extract has no locality field, we can still read codes and hours, but we cannot name streets. We will say so in the quote.
Fees
From RM 4,200 for a single week at one hub. Street-level clustering needs a locality field; without it the fee is lower and the note is narrower.
Next step
Send a one-week extract (or a description of what you can export) and we will say whether a study is worth running.