Dashboard

Marco is the outlier, and consistently high rather than random -- he scores generously. That is coachable and it is also a reason to look at whose work he happened to be assigned.

Amber is the median analyst. Red is Marco's distribution sitting to the right of everyone else's.

Window

8:00am to 8:00pm

Timezone

Each lead's own local time

Weekends

Saturday only, 10:00am to 5:00pm

Held now

0 leads outside window

Enforced in the queue, not by you. A lead outside its window is never handed out, so there is nothing to remember and nothing to get wrong at 8:01pm.

DNC Shield · federal

not connected

DNC Shield · state

not connected

Internal

2,914 · updated continuously

Text opt-outs

STOP replies · automatic

Numbers change hands. Someone who agreed to be texted in March may not own that number in September, and the person who owns it now never agreed to anything. Checking the FCC's database before messaging is what earns the safe harbour if you get it wrong anyway.

Recording turns an operator's word into evidence -- but roughly a dozen states require every party to consent, and California is one of them. An unannounced recording there is not proof, it is a second problem.

Every consent record stores which version was read aloud. Changing the wording never rewrites what somebody actually agreed to.

Carrier brand

not registered

Campaign

not registered

Phone number

none purchased

STOP handling

tested end to end

Calls work today. Texting does not send until the carrier approves the campaign -- a review queue nothing here can shorten. Opt-ins captured now are waiting for it.

Component

OMNILINK - CRM Connector · upstream

State

queued, not sending

Waiting

41 payloads

QA is the only role that can score, and the only one that cannot work. It is not a supervisor with less access -- it is a department with different access, which is why neither column is a subset of the other.

Last import: Sep 1 · 1,204 rows accepted, 22 rejected as unreachable line types.

Every field is marked unknown rather than guessed. A mapping that looks configured but points at the wrong field is worse than no mapping -- it fails silently and you find out from a customer. do_not_call is bidirectional on purpose: a DNC anywhere must be a DNC everywhere, immediately.

Abandonment

not measured -- manual mode

Cap

3% per campaign per day

Ring time

15 seconds / 4 rings minimum

One keystroke drops a pre-recorded message and moves to the next lead without waiting out the greeting. Against roughly forty-eight no-answers a shift this is the single largest time saving available anywhere in the console.

Every drop identifies the caller and the business by name and gives a number to call back. A voicemail that does not say who left it is the fastest route to a complaint.

Recording turns a consent record from an attestation into evidence. It also creates its own exposure: some states require every party to consent, so an unannounced recording there is a liability rather than proof.

Numbers

0 purchased

Caller ID name

not registered

Branded calling

not set up

Attestation

not verified

Local presence lifts answer rates and raises your spam-flag risk at the same time -- carriers watch for one number pattern dialing many area codes. If you use it, use few numbers and rotate them slowly.

The first five map onto Odoo's own HR models cleanly, so they are not guesses. The last two have no natural home -- certifications gate queue access here, and an OSHA incident is a legal record with its own retention. Both probably stay in A-TEAM and get referenced from Odoo.

Meal and rest rules follow the strictest state anyone on the roster works in, not the state the company is registered in.

A telemarketing or call-centre operation may fall under OSHA's partial exemption from routine injury-log recordkeeping -- it depends on your NAICS code and whether you had more than ten employees at any point in the last calendar year. That is a determination to make on paper, once, with the codes in front of you. Not a guess.

The bottom row has no OSHA standard behind it and it is still where every real injury in a call centre comes from -- hearing, wrists, backs, eyes. Training on it is worth doing because it works, not because someone requires it.

Reports can be anonymous, and an incident cannot be edited or closed by the person it was reported about. A safety report a manager can quietly alter is not a safety system, it is a liability with a form attached.

Nobody Left Behind is the most useful contest on this list. It pays the whole floor only if the weakest operator improves, which turns your best people into coaches for eleven days without anyone being told to mentor.

A bonus is a payroll record, not a game score. It has to be traceable to the outcomes that earned it, disputable on the same path as a QA score, and it has to land in payroll as earnings.

Metering exists before pricing on purpose. You cannot price a seat-plus-usage product until you know what a real client actually burns in a month, and this is month one.

Nothing on this schedule runs today. Writing it down is the first half; a job that actually deletes on it is the second, and the second half is what a regulator asks to see.

A California business owner can ask what you hold on them and ask you to delete it. Today that is somebody writing SQL against a live database, under a clock.

When a claim arrives, everything touching that number has to stop moving -- no deletion, no retention job, no edits -- and somebody has to be able to prove it stopped. There is nothing here that does that today.

Publishing a new consent script version does two things automatically: it marks every operator uncertified on that script, and it holds them out of queues that need it until they pass. A wording change is a legal change, so it cannot be a memo nobody read.