The document summarizes well the issues around time tracking. We have tried many times over the years: now we have a more or less stable situation, but one wonders about the accuracy of the data.
As the paper suggests, if you have a very detailed concept, using it becomes a chore. If it is too simple, it might not reflect
reality, and we might be better off going back to assumptive cost allocation: we
kind of know what a team is doing, e.g. what applications they might develop or support, etc. This approach would also apply to release trains in SAFe.
After many years of debating the topic, we came up with a set of activities per team. If I have to judge by how I use it, there is little value in the exercise: either it's a terrible overhead, or we just use it to get away with the bare minimum. This is well reflected in the paper.
Our DevOps teams use a very straightforward method: all their labor costs go directly to the BU that sponsors them. No details, but no headaches either.
------------------------------
Guillermo Cuadrado
Senior Technical Manager
Amadeus Data Processing GmbH
Erding
+49 172 8696 790
------------------------------
Original Message:
Sent: 05-13-2022 16:38
From: Bill Kasenchar
Subject: Time Tracking - Pro and Con
As with any other data source you need to evaluate the business need against data quality as categorized by good, better, best. I have done extremely detailed time tracking by task, tagged to service, project, and application meta data – updated nightly from Planview to Apptio via API. And then thrown it all away…
There are a couple of instances where detailed time tracking is important and valuable
- Time and material chargeback by hour (generally app dev or support services by a shared services organization)
- Separating capitalizable labor from operational labor
Even for these instances – Keep it simple!
For most other use cases time tracking is usually more burden than it is worth and has it own potential accuracy issues.
First answer these two questions…
What business decisions or budget decisions are dependent on granular and accurate time tracking?
How much deviation in cost is there between super accurate, granular time tracking vs estimation by percentage of FTE?
Whether or not you do time tracking, you still need to allocate labor, so there needs to be an understanding of your fully burdened internal and external labor costs. For privacy reasons, fully burdened labor costs should not be by named resource. It is better to have an (annual) average rate by role and allocate that to the associated products, services. This approach also solves the maintenance issue as resources are added or removed from tasks, projects, products, etc. Resources only need to be mapped to a role. This can be a single schedule of role rates or there can be one for internal labor and one for external labor – if they differ significantly.
Attempting to reconcile external labor time tracking to contractor invoices from the GL is fraught with issues. Primarily, there is latency between time tracking and invoices from contracting vendors. If travel expenses are involved, it makes it more difficult. It will not tie out and time tracking could be 30, 60, 90 days ahead of what is in the general ledger and at times this information crosses fiscal boundaries.
In most use cases a logical allocation of FTE cost by role to a service, product, application, etc is sufficient. For example, if the service owner for storage has 5 FTE working for him/her, is time tracking going to give that person any better understanding of the demand on those individuals? Is it going to change the overall cost per GB for storage? Are they spending significant time on tasks that are outside the scope of storage and back up services? Are those individuals funded outside of that infrastructure department?
Similarly with tickets or stories it is difficult to come up with a time spent on per unit basis that accurately reflects effort and cost. The relationship between how long a ticket is open and how much effort by skill level of resource is vague and varying. I have found that for cost use a simple allocation method but for continuous improvement use ticket count to know which products or services need the most attention.
Anyone that has heard me speak or attend one of the TBM classes I've taught, would have heard me talk about the distinction between a cost model and a spend model. Related to time tracking, the spend model (money out the door) ties out to the general ledger and is using the contractor invoices as they are incurred. This model would also be used to show budget variances. However, the cost model shows the cost of services, products, etc inclusive of labor where labor in the cost model is based on role rate regardless of the mechanism of determination (time tracking, % FTE, hybrid, etc). This is what will be used for showback/chargeback to the clients. There will be a variance between cost model and spend model with this mechanism so take the variance into account when adjusting role rates for the next fiscal period – improving accuracy and minimizing the variance YoY.
attached is a deck describing the justification to dramatically reduce the time tracking burden
------------------------------
Bill Kasenchar
Principal
Higher Logic TBMC Community Test Account
Philadelphia PA
2158980001
Original Message:
Sent: 05-11-2022 11:28
From: Christina Pastella
Subject: Time Tracking - Pro and Con
We rely heavily on time tracking in ServiceNow to projects, enhancements and operations. However, we do not require a full 40-hr time card from all resources. Staff Augmentation resources and non-officer FTEs are required for full time cards. However, officer resources and vendor managed service teams are required to only report project time. We use datalink to bring in time data directly, and use it to calculate actual time worked to projects in PFP, and also created reports in CT to auto-allocate vendor invoices to projects and operations based on time data in SN.
We are encountering difficulty in keeping the labor pool fresh and updated with correct rates in Apptio to support our reporting - right now it's a manual effort, as we do not have a system that ties resources to contracts to rates in one place in real time. Researching options there and am hoping for more responses to this thread to gauge if what we're doing is most efficient..
------------------------------
Christina Pastella
SENIOR IT FINANCE MANAGER
Tokio Marine North America Services
Bala Cynwyd PA
Original Message:
Sent: 05-10-2022 07:50
From: Chris Curtis
Subject: Time Tracking - Pro and Con
We are in the process of eliminating the majority of time entry with our strategic shift away from projects to products. Our new methodology is going to leverage the JIRA ticketing information to perform a cost allocation to the work performed based on the labor, contracts, and non-labor in cost centers. We are in deep discussion on whether the basis should be the # of stories or the # of story points per story/work item/task in JIRA. Historically we were heavily reliant on time entry to track spend on projects and to allocate costs.
Chris
------------------------------
Chris Curtis
Director - Change Management and Data Analytics
Freddie Mac
McLean VA
410-215-4952
Original Message:
Sent: 04-30-2022 21:55
From: tim pietro
Subject: Time Tracking - Pro and Con
Community,
I've been asked a lot by the TBM Community if time tracking is still being used today across IT Organizations. For some companies, you can't track the time of FTE's. In my old-old-old IT life, I would track via excel and would ask service teams to provide me their efforts by a % by service offering (not service).
As of today, where do you stand with time tracking for projects and TBM integration?
Tim
PS: Thanks in advance for the feedback.
------------------------------
tim pietro
Practice Partner
Wipro - Partner
Norcross GA
(650) 793-5810
------------------------------