LTB 678.16 - National Terms of Reference : Predicted Time of Delivery Data Background Work

Post by **TrueBlueTerrier** » 16 Nov 2016, 16:37

LTB 678.16 - National Terms of Reference – Predicted Time of Delivery Data Background Work

No. 678/2016

Wednesday 16th November 2016

National Terms of Reference – Predicted Time of Delivery Data Background Work

Dear Colleagues

Branches and Representatives will be aware of the number of projects relating to delivery related trial activity as Royal Mail bid to compete in the Parcels market, and improve the service provision for customers by improving the final mile experience for the customer through digital/mobile applications.

Aspirations within the current Commercial sphere include providing improved online information/SMS notifications to customers regarding the arrival of and delivery of items at the doorstep and giving greater in flight options around delivery preferences such as nominated day, time or different address.

As a result we have agreed to undertake a Data Validation Study which will explore the feasibility of RMs aspirations to offer a Predicted Time of Delivery to the customer.

The study will continue to take place at Cardiff North Delivery Office and will be extended to include Burton-On-Trent DO and Cambridge DO and will run from Mid November 2016 – Mar 2017 inclusive.

The study will take place in the following manner and explore:

how an algorithm can be created within the RM environment to predict when a parcel may be due to arrive with a receiving customer
which data sources within RM are required to provide the best input for the algorithm to provide an accurate delivery timeslot (e.g. routing data, historical scan data, locally held data sources)
the role of start times within the algorithm for generating accuracy for a Predicted Time Window
how best to structure the timeslot provided to a customer (e.g. if a 1 hour ‘buffer’ is set around the predicted time an item is due at a delivery point,)
what operational scenarios may need to be excluded from the algorithm method of prediction (e.g. adhoc routes, timed delivery, etc)
how all route types can be catered for in the algorithm (e.g. HCT, shared van, singleton van)
what ‘on-the-day’ operational scenarios exist and how the algorithm can cater for these or where alternative solution design may be required (e.g. additional technology or process changes) to ensure accuracy in what is shared with the customer
what the change impact of introducing Predicted Time of Delivery to a unit would be (e.g. managerial tasks, OPG information needed ‘on-the-day’ to support prediction accuracy
Draw on RM data sources from differing units to understand how well/poorly the latest iteration of the algorithm is performing
Face-to-face discussions with frontline employees

It is important to stress and bring to the attention of Branches that through the Study we have agreed the following commitments in relation to individuals:

The discussions and data from this work is not for performance-monitoring purposes
Individuals are not required to take part in discussions with the project team. However assistance with understanding the characteristics (timings, volumes, patterns) of a route are appreciated
It is confirmed that any data or information gathered for the purposes of the trial will not be used for performance-monitoring purposes. In addition the information is for the purpose of future customer product enhancement only, and therefore will not to be used for any individual conduct or work study purposes

The intended outcome of the study is to inform senior Commercial stakeholders as to the feasibility of providing a Predicted Delivery Windows attribute on parcels within RM, requirements and solution design that is requested of IT suppliers to support the provision of the service and whether there are other opportunities within the provision of data to exploit commercial opportunities, which will form part of the National RM/CWU Growth Forum discussions and future talks on Future Design/Pipeline.

We have agreed that other delivery offices may be brought into scope but prior notification would be provided to the national CWU before any on-site discussions with frontline employees take place, to ensure we have made Branches aware and units selected are suitable.

In addition the method/system of how the algorithms predictions and associated data are generated will be shared with the CWU nationally in order to ensure there is joint understanding of these factors and how this will be cascaded into Deliveries in the future.

In agreeing to the above commitments Royal Mail have stated that they intend in the future national talks to enter negotiations on how the data and information could be used. We have reserved our position regarding this at this moment in time, however it is important to ensure Branches and Representatives are aware that this aspect is something that will form part of the future discussions and will be subject to national agreement.

Any queries to the content of the above please contact the Outdoor Department reference 530

Yours Sincerely,

Mark Baulch

CWU A/Assistant Secretary

Post by **IloveMYredTROLLEY!** » 16 Nov 2016, 17:11

This is never going to work very well, park and loop van duties are regularly forced to change the loop order to get all 1pm Special Deliveries done on time, while HCT duties are often delivered in wildly different orders (not just by different OPGs, but how busy a day is and how stupidly late their replenishment bags might be available etc.).

steve1873 · Post by **steve1873** » 16 Nov 2016, 17:47

For this to be anywhere close to achievable, then Royal Mail need to do away with the 1pm deadline for Special Delivery and just have it as guaranteed next day, with a further premium added perhaps to ensure delivery by a certain time (a special duty could deliver these all for an individual office). The amount of times I have to swap around loop orders and do an change things around, simply to get to a certain area prior to 1pm, means that no matter what "algorithm" they employ, it will simply be impossible to predict with any accuracy a when a parcel will be attempted.

Also, what is with this "giving greater in flight options around delivery preferences such as nominated day, time or different address"?

I think there better be VERY clear limitations as to the different address options once commencement of delivery has been made!

postslippete · Post by **postslippete** » 16 Nov 2016, 18:33

What planet is this guy on??

A stereo-typical project manager with the best of intentions for improving our business with our customers BUT without a clue of how our duties are being covered on a daily basis.

Its not just the fact that posties regularly alter their routes and loops due to getting their special deliveries delivered before 1pm. What about when posties often drop a few loops on busy days and instead of the customer getting a predicted delivery time of 10.30am, it arrives at 4.30pm by someone else delivering instead

Post by **fb1969** » 16 Nov 2016, 18:56

Dear Royal Mail Project Team

I can predict the delivery time of any item to within 5 minutes.

But I can only give the predicted delivery time approximately 5 minutes before I get there.

Hope this helps.

worktotime · Post by **worktotime** » 16 Nov 2016, 19:06

another great idea from the powers above who don't do the job everyday and haven't a clue , easy option put parcel dutys back in like the old days

, if they look at the other companys who have timed deliverys on parcels they don't have specials and they don't deliver mail , come on its not rocket science

steve1873 · Post by **steve1873** » 16 Nov 2016, 19:13

I think we should give them a 4 1/2 hour window!

k979aaa · Post by **k979aaa** » 16 Nov 2016, 19:30

JUST TRY YOUR BEST

worktotime · Post by **worktotime** » 16 Nov 2016, 20:02

its doable.

DB1100 · Post by **DB1100** » 17 Nov 2016, 17:57

I sense clusterfeck

Newbie001 · Post by **Newbie001** » 17 Nov 2016, 19:19

Simple, take any specials that cause you to cut off or change order of loops. And get linos to do them. Can't see that happening though.

smileski · Post by **smileski** » 17 Nov 2016, 20:37

'The discussions and data from this work is not for performance-monitoring purposes'

What worries me is, just how naive have the CWU become?

datasaint · Post by **datasaint** » 19 Nov 2016, 16:32

Abandon van share. Have solo vans/rounds. Give a one hour window for deliveries. Would easily be achievable. Even on the heaviest days you aren't going to extend much more than 1 hour. And you can easily break off a loop to meet a scheduled delivery if you absolutely had to.

wannabeek · Post by **wannabeek** » 19 Nov 2016, 16:48

1 hr window every day? Really? During this year my first letter has varied from approx 8:15 am at the earliest to 9:45 at the latest. Last letter of the day has varied from about 11:30 to 17:30. Also, the loops are very regularly done in varying orders depending on specials and road traffic - if I was obliged to follow the frame religiously then the last letter would have been even later.

Post by **TrueBlueTerrier** » 19 Nov 2016, 16:49

norm wrote:Abandon van share. Have solo vans/rounds. Give a one hour window for deliveries. Would easily be achievable. Even on the heaviest days you aren't going to extend much more than 1 hour. And you can easily break off a loop to meet a scheduled delivery if you absolutely had to.

thats the next plan