0

I'm a statistical modeler at a large bank in a highly regulated area. I work in a team of about twelve people divvied up between three managers. Right all of our statistical models are developed using a simple waterfall style methodology which where the models go from "Data Collection" to "Methodology" and then to "Specification and Testing" among other milestones.

The current management keeps setting target dates that are far too aggressive for model development in this framework, as a result most of the modelers are working long hours, weekends, etc... I'm taking a break from writing up documentation on a Sunday morning to type this. The current set up for development came into being seven months ago and since then our turnover rate has skyrocketed. In a 12ish person team we've had five people quit or transfer out of the department in seven months. As a result I'm the only non-manager in the group who was with the bank prior to the re-organization. I'm also contemplating leaving the department.

I strongly suspect that the issue is that the current head of the team and his managers have no idea how to set deliverable dates for this kind of work. All of managers come from an analytics background and seem to have no experience with model development. I wrote a memo several months ago encouraging the team to explore an agile development style but this was largely ignored, we are not co-located which makes adoption of pure agile methods difficult in any case.

How do I tell the managers to get it together on the project management front and either set sensible milestones for our waterfall tracking (using something like critical chain?) or move to Agile?

P.S. for this kind of development work it's extremely difficult to forecast how long development will take until you've done at least a couple weeks of preliminary work (getting data/cleaning data/etc...) .

  • 1
    What other consequences are taking place? What happens when the team is late in delivering something on the aggressive schedule? Are there quality issues in the team's work? What is the degree of cost impacts due to weekend and overtime work? Is anyone getting trouble for being late? – David Espina Aug 7 '17 at 10:18
  • 1
    Have you done anything else but the memo regarding the situation? Have you escalated risks due to the aggressive schedule, e.g., being late, turnover exacerbating being late, quality jeopardy, low morale, OT costs? Are you briefing management weekly, showing "red" on performance metrics? – David Espina Aug 7 '17 at 10:21
3

I strongly suspect that the issue is that the current head of the team and his managers have no idea how to set deliverable dates for this kind of work.

That isn't a suspicion, it is a hypothesis with preliminary support. Quite frankly I suspect it is like a hypothesis that water is wet; it is a no brainer.

Every time the schedule is set by people other than the people doing the work, that schedule is a work of speculative fiction. It is based on hopes and dreams ; it is an incarnation of everything we know about bias

In your shoes, I'd leave. But if you have a greater appetite for risk than I do, I think you have the beginning of your answer in your question.

You have the opportunity to present to your management a significant cost savings opportunity.

You have a 12 person team that has had 50% turnover in less than a year (Or more - I'm not sure I'm reading the statistics right). You'll need to refine my estimate but let's work on a 50% turnover rate projected forward. the cost to hire and fire varies by industry and location, but it is easily in the tens of thousands of dollars (at least in my industry). Let's assume that the average turnover is 10% and your team is 50% above that - so the cost of doing nothing is approximately 75K/year. That is a large enough impact to get the attention of most managers - "What if I gave you 75K/year to spend as you like? And what if that resulted in significant increases in long term quality of the data models?"

In my humble opinion, the fix is trivially simple. Managers should never set delivery dates. Project Managers should consult with subject matter experts; SME set the schedule and project managers report on the schedule. Schedules that are set by the SME are going to be reliable. SME will work the schedule because they know it is fair to them. Turnover will decrease, the resulting data models will be higher quality and more reliable (nothing developed on nights & weekends is truly reliable; that kind of death march is a guarantee for errors).

There are a variety of techniques to improve the scheduling even further, but none of them will work when the accountability for the schedule is divorced from the responsibility for setting the schedule. That's a no brainer. I think you can probably save the bank even more money by adopting something like Evidence Based Scheduling.

Personally I think you should get a % of the 75K you save. Feel free to kick back a percentage of that to me.

2

A couple of additional points for a discussion with management:

  • point out the 40% turnover rate under the current setup

  • note the time spent in replace staff (hiring campaign + training)

  • observe the overtime people are working; its known or plausible correlation with turnover rate and thus risk of further turnover; point to studies on unsustainability, burnout, loss of productivity

  • ask what is driving the aggressive schedules: there may be something important going on that isn't filtering down to your level, and if you don't take it into account, you will have small chance of swaying them.

Something else I might do before this conversation is make my own estimates for one or two rounds of deliverables. I'd estimate in work hours, then compare to the actual work hours spent including nights&weekends to meet the deadline. If my estimates were more accurate (as I gather you expect yours to be), then I would suggest comparing estimation methods, and let them go first; then show mine, & convert them into more realistic deadlines.

Good luck!

0

From personal experience,

This is not a 'hey you are wrong, lets do it this way' kind of easy discussion in my experience. It could take longer. Directly sharing benefits and cost savings on following a process may not help because you probably are not the boss who can fire them, and so they probably won't give a damn about your suggestion of benefits to change. They have to be interested and participants to the change.

  1. Create a Forum you should make a case for a project management office PMO weekly / bi-weekly / at least a monthly call.

  2. Create an exchange and discussion even before plating the concern you have. Try to confirm different problems PMs have, and if they have thoughts on improvements.

  3. Plate the concern, benefits / risks of change and ask their thoughts. Next step is to get them to agree or convince some changes are required..

  4. Closing - Document the changes - one of them could be a PM methodology to follow. Document the benefit/risk of the changes, Document the RACI for the change in the process.

0

Nothing motivates change more than ongoing issues, so long as those who would sponsor the change attributes those issues to the right causes. The knee-jerk, easily identified cause is to blame people, so getting your bosses to see the other drivers is key.

I asked about what your other interventions have been, including what risks have you escalated and documented that ended up coming true. Aggressive targets lead to quality, cost, and morale / turn-over jeopardy but it also cures pessimistic estimating from the team and Parkinson's Law / Student Syndrome, both of which add costs but no value.

I do not see an issue with management dictating targets so long as it is based on estimates provided by the team. If the target is aggressive, management may be signalling something around pessimistic estimates they typically receive, are imposing a challenge to the team but have no real expectations of meeting those targets, may be resolving other issues that impact the business of which you are not privy. But your team would need to simply and formally escalate the threats as you see them and then mitigate the best you can.

But introducing a change that would cure the threats / issues you're facing will only occur if your bosses perceive issues. The question remains, what are their complaints? Who is getting beat up for misses or financial impacts? Who is paying the price of staff turnover. You're case for change will need to be centered on the answers to these questions.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.