This article concludes our first major series, the 7 Steps to converting actuarial models. Read that overview post for a quick summary and links to everything that’s come before
As a conclusion to a series, then, you might think we’d be talking final specifics about conversion projects. In fact, we’ll be talking more about the future. We’ll discuss how to set yourself and your department up for greater success in future conversions, or similar projects. [Obviously, nobody can guarantee success. But can you increase the chances? Absolutely.]
Why? Because it is likely this conversion isn’t the last time you’re going to go asking for approval to make a major change.
And when you do, many of the same people who approved this project will be on the team to approve the next one, as well. You want to be able to show them you’ve delivered a positive return on this investment.
Remember – as actuaries, your role is much greater than simply the technical aspects: calculations of reserve balances and roll-forwards, evaluation of what-if scenarios, etc. The actuary’s responsibility is to create value through risk analysis and recommendations. If you forget that, you will be relegating yourself to a lower-tier of responsibility: data management, hardware management, or simple processing.
With that lower-tier comes lower visibility, lower authority, and, ultimately, less job satisfaction.
Not to mention lower prestige for the actuarial department and lower trust in the results you’re delivering on the regular.
You didn’t get into being an actuary for the databases. You most likely got into it for the opportunity to work on “interesting, compelling problems”. Unless you have a good way to demonstrate that you (and your department) have solved those compelling problems for your employer, you won’t be given additional opportunities to do the same in the future.
Leaving you stuck at the “code monkey” type of role, and giving data scientists, software engineers, and business analysts the chance to have all the fun.
So what can you do to ensure that you’re working at the top of your credential in the future? Easy – demonstrate the return on investment in a project.
The current example is to show how much value you have created for the company by converting your actuarial models to a modern system. Yet the principle holds true for any project: review the work done, make an appropriate comparison, and use insights gained to position yourself and your department for future success. How to do that? Follow these tips below and you’ll be well on your way to a compelling argument.
Keep a log of new practices and compare them to prior ones
This relates directly to the things you wanted at the start of your model conversion project.
You should be asking questions like these: What did we want to get out of that conversion? What did we actually get? Do we have what we wanted? Is it more or less than anticipated?
If it’s not satisfactory, why? Is it inefficient application of the system? Unreliable data? Lack of professional knowledge? Or system capacities that were over-promised and under-delivered”
And, importantly, What else could we be doing differently?
You will most likely only be able to answer these questions if you’re regularly keeping track of what’s come up since you’ve put the new system into production.
Unless you have some data, you won’t be able to tell how well you’re doing. Anecdotes are fine for complaining about your boss during happy hour, but they’re not evidence.
Using anecdotes as evidence is subject to being significantly influenced by one or two loud voices. They may not represent the entire spectrum of experience.
Evidence is that comprehensive record of what you were hoping to get and whether you did. In order to make your work product of the future even better, you need to know where you stand. As a result, keeping track of outcomes and comparing them to expectations (that old actual-to-expected calculation that is the staple of early actuarial work) is vital.
Survey associated teams to see what changes have emerged in their practices
By now you should have new measures of your various outputs. Maybe you can point to a shorter turn-around time for financial close process, reduced pricing cycles, etc. Or you may be able to produce regular analyses without asking other teams to run reports. Maybe you now have risk management solutions that are clearer or are more actionable than they were before.
Here’s an example: a client of ours modernized their actuarial software and was able to reduce their pricing cycle from 2 weeks to 4 hours. That 95% reduction means they have the opportunity to vastly improve many of their other processes as well.
The point is, in order to make your own informed decisions, you need to know, not guess, what other people around you are seeing. So, you probably should be asking other departments what they are getting out of your modern actuarial department and its modern actuarial system.
To get you started, here are questions to ask of five different audiences that are likely to be impacted by the improvements you’ve made:
|IT||Do you find it easier to manage our permissions, data storage and retrieval requests, and compute resources? If so, can you quantify the amount of ease? Time savings? Hardware costs?|
|Underwriting||Do you find that enhanced reporting and analyses we provide to you enable you to perform your decision-making with more clarity, speed, or effectiveness?|
|Administration||Do you find it more or less burdensome to provide us with policy extracts or error resolution each period (week, month, quarter, or year) for the various tasks you support?|
|Finance||Do you find greater accuracy, efficiency, or actionable decision-making recommendations from the information we provide about financial elements? Do you see reduced errors or shorter turnaround time on requests? Do you notice your own decisions are made better or with greater confidence because of the improved process for creating ledger values?|
|Actuarial Management||Do you find that the reporting explanations and risk management recommendations we provide are more actionable, detailed, or valid to making business decisions?|
Obviously, this isn’t going to cover everyone. And you could expand these questions almost to no end. Certainly business consultants could help you design a broader feedback program. But this list should get you started on systematizing that feedback process.
Revisit your cost comparisons
Now that you’re done with the technical aspects of model conversion, you can take some time to look back and see just how much you didn’t know when you started and have learned along the way.
Because of the fluid nature of these projects, you may end up with a setup that includes a structure very different from what you anticipated. Or, you may now have more context to compare between systems: flat fees, per-license fees, technology support, etc.
You’ve probably made the case for conversion a while ago. What’s new? What’s different now?
What is the length of time that this conversion has gone on compared to what it was planned to be? Plug those into your break-even calculator and see where your ROI lands.
Once you have that, you’ll be able to determine just how much “transformation” and “streamlining” you’ve been able to accomplish.
Also, if you’re just finding this now and think you might need a tool for comparison, we have a spreadsheet that might help you with categories and quantification. Send an email to info@slopesoftware and request the “Cost of Software” spreadsheet benchmarking tool.
Which means that your retrospective look doesn’t just compare your costs now to what they were before the whole conversion. Remember, it’s likely that you’re also doing a lot more (or doing it more efficiently) than you were back then. It’s not quite apples-to-apples comparison.
You have to find some kind of level basis. And that level basis is usually “what would our costs look like now, but using the old regime?”
In P&C insurance this is the “on-level premium” analysis. You can add inflation if you like, but it’s probably not going to have gone on long enough that inflation makes a material difference.
Essentially, though, what you’re really trying to do is develop a fair comparison. The best way to do that is to revisit that cost comparison you made before, and update with all the new information you have learned since then.
Then, run the same cost analysis under the new system, and you can find that reasonable basis that allows you to evaluate elements that were so dissimilar before.
Bring it all together
Now you’re combining the work you’ve done under the first 3 tips here to make the case that moving to this new system is worth it.
Look at your processes, how they impact others, and the costs involved. Perhaps you write a summary memo, or a report, or put together a presentation for your department or even the profession at large. This would be a good way for you to synthesize all your learnings and ensure that you continue to reinforce them as you go forward.
And remember, you’re always keeping an eye on the future. It’s not just now, it’s also what does the actuarial system look like 5 or 10 or 15 years from now?
If you’re going to need another change to incorporate whatever new situations arise later, doing it well this time will ensure you are well positioned to make a good decision then.
What if you didn’t get what you expected?
Every data point helps advise the future
Suppose you were promised that you would see 25% reduction in time to complete quarter-close processes. If it’s only been 15%, or 10%, what then? Do you just chuck it all and go back to the drawing board?
Here is where we point out the Actuarial Control Cycle. Really, this is just the business cycle, in which you evaluate a current position, identify opportunities, suggest refinements so that you adapt to changing circumstances, and move forward.
Which may include another round of model refactoring so that you can start to leverage whatever functionality you thought you were going to get, but which got delayed in the push to converted.
Don’t overweight sunk costs
We need to revisit the idea of “sunk costs”. All actuaries are familiar with the theory that amounts spent in the past shouldn’t influence the decisions about spending in the future.
Theory is all well and good in theory, but we know that in practice many actuaries (and finance professionals) are loathe to recognize sunk costs as lost.
They often allow those recent experiences to influence their later decision-making. Intellectually, we know it’s not a good way to go about making decisions. Yet we still do it. Perhaps you should try not to in your next evaluations.
That is, the amounts you spent in the past should not be a factor in the decision on future investment. The decision-making process, however, absolutely should be a factor.
If you made bad decisions, and are asking for continued bad decision-making, it would be unwise to keep approving your ideas. But if you have made a good decision in the past, and the results just didn’t work out (maybe they were at your 15% likelihood scenario, which, remember should happen every once in a while), you’ll have more credibility.
The only way you can know this is to actually measure the output and compare it to what you expected to see.
That’s how you’ll demonstrate good decision-making and instill confidence in your ability to execute next time.
Plus, this conversion project is potentially going to impact your reputation (your being the collective “you” as the actuarial / finance / underwriting / IT world).
You want to do everything you can to not only validate the decisions you made before, but also set the precedent that those who are trusting you with the time and financial investment are going to be repaid well by that trust.
Keep the future in mind
All of this points to the future of actuarial work. In the past, it might have been enough for the actuaries to just know all the reserve methods inside and out, and make a career of quarter-end financials.
In the future, though, actuaries can no longer be just “databases with faces”. That’s not going to be enough. It’s not a robust enough view of the world or modern actuarial work.
It’s limiting. It’s the old way, and if actuaries don’t advance their practice, they’ll be left behind.
Actuaries can and should probably consider themselves more like an “insurance system engineer” than an applied mathematician. Why?
Because society continues to build new technology, such as robotic process automation and business intelligence software. What is the point of all that technology, if we don’t use it to actually make our actuarial work product better too?
It shows up in automating (or making automate-able) many of the old processes that actuaries used to spend their time on.
Freeing them up to actually, you know, be actuaries again. By delivering actionable risk analyses and recommendations.
Often, they can only do that if they have enough time to analyze the results. And how do actuaries ensure they have enough time? By simplifying processes and incorporating modern actuarial software to do a lot of the work for them. I.e. “modernization”.
And thus we get back to thinking about the next project.
It’s a perpetual cycle. One which, if done well, can lead to incredible results.
Quick summary: revisit your work, make an evaluation, and allow that to inform your future decisions.
There are so few of these events happening that it’s very limited on how long a model conversion “should” take and what steps are involved.
With such a small data set, every piece of information provides significantly greater marginal information about how this process goes.
Because remember – it’s likely this isn’t the last time you’ll ever go through a model conversion. Either the current system will become obsolete or you’ll move on to another company. Or maybe your company will buy a block of business and need to integrate that model into your current workflows. [All of these are good reasons to look for a new system.]
We shouldn’t believe that one system is going to be sufficient for the rest of time. It just doesn’t work like that.
The point is, you want as much information to refer to during the next transition, so you can continuously refine your analyses and constantly improve your processes.
When the next conversion comes, you want to be prepared for a good transition. Like so much of actuarial work, successful model conversion is much more likely when you make informed decisions.
And you can only do that when you have the appropriate data. So get it!
When you do, you can make the next time an experience you look forward to, rather than a chore.
We’ll see you there.
Next Time: there isn’t any more! That’s the end of this series. See it from the start.
Maybe model conversion shouldn’t be so infrequent: Find out about Hidden Benefits of Model Conversion by sending an email to email@example.com.