From OT Assessment to Action: Building a Practical 90-Day Improvement Plan
Most OT security assessments produce a competent report that changes nothing. The findings are accurate, the recommendations are sound, and twelve months later the same assessor could issue the same document. This is not usually a failure of the assessment. It is a failure of the step after it, which almost nobody plans for and which is where the value either gets realized or does not.
The gap is structural. An assessment produces findings organized by risk. An organization executes work organized by budget cycle, outage window and who has capacity. Those are different shapes, and nobody converts between them. This article is about that conversion, and the first 90 days specifically, because that is the window in which a report either becomes a program or becomes a PDF. If you have not run an assessment yet, our OT risk assessment walkthrough covers that step.
Sort findings by what it takes to fix them
Risk rating tells you what matters. It does not tell you what you can do this quarter, and treating a risk-ordered list as a work plan is the single most common reason nothing moves. Re-sort the findings into four buckets.
|
Bucket |
What it contains |
Typical timeline |
|
Documentation |
Evidence that exists but was never written down, or records that were never kept. Risk acceptance, approvals, inventory gaps. |
Days. No outage, no budget. |
|
Configuration |
Changes achievable on live systems: accounts, rules, logging, default credentials. |
Weeks, inside normal change control. |
|
Outage-dependent |
Anything requiring the process down: segmentation changes, device replacement, firmware on critical assets. |
Next planned window. |
|
Capital |
New hardware, platform projects, equipment replacement programs. |
Next budget cycle, often the one after. |
The first two buckets are your 90 days. The third and fourth are the plan beyond it, and separating them early is what stops the whole program stalling behind one capital item.
|
Why the documentation bucket goes first It is tempting to skip documentation findings as administrative and start with something technical. Do the opposite. Documentation items need no outage, no budget and no procurement, they close a disproportionate share of audit findings, and finishing them inside the first month builds the credibility that funds everything else. A program that reports real completed items at the first steering review gets its capital request taken seriously. One that reports "in progress" against everything does not. |
Every action needs a name, not a department
An action assigned to "OT team" or "IT security" will not happen. An action assigned to a named person with a date will, or will visibly not, which is almost as useful because it surfaces the real constraint. Three fields per action and no more: who, by when, and what done looks like.
That third field is the one usually missing. "Improve network segmentation" has no completion state and will be open forever. "Conduit between Level 3 and Level 3.5 enforces an explicit permit list, any-any rule removed, change record filed" either happened or did not. Write the acceptance criterion at the point you assign the action, not at the point you review it.
A 90-day shape that works
- Days 1 to 15. Convert findings into the four buckets and assign owners to everything in the first two. Agree the acceptance criterion for each. This is a workshop, not an exercise in a spreadsheet.
- Days 15 to 45. Close the documentation bucket. Risk acceptance records, approvals, inventory corrections, policy ownership and dates. Largely desk work, and it should finish.
- Days 30 to 75. Work the configuration bucket through normal change control, overlapping with the above. Default credentials, dormant accounts, logging, vendor access cleanup.
- Days 60 to 90. Scope the outage-dependent items against the next window, and build the capital case for anything in bucket four with costs and consequences attached.
- Day 90. Report what closed, what is in flight, and what needs a decision. The decisions are the point of the review.
Measure closure, not activity
Three numbers make a steering review useful. Findings closed against findings raised, which is the only completion measure that matters. Findings closed by bucket, which shows whether you are making progress or only making progress on the easy ones. And the count of actions with no named owner, which should be zero and is the leading indicator of a program about to stall.
Resist reporting effort. Hours spent, meetings held and documents drafted are inputs, and a program that reports inputs is usually one that has stopped producing outputs. Our guide to OT security metrics for the board covers the reporting layer above this.
What to do with findings you will not fix
Some findings will not be remediated, and that is a legitimate outcome rather than a failure. Legacy equipment that cannot be patched, a control that would cost more than the risk it addresses, a recommendation that does not fit your operating reality. The requirement is that the decision is explicit: documented, with the reasoning, accepted by someone with authority, and given a review date.
An assessment with 40 findings where 25 closed, 10 are scheduled and 5 are formally accepted is a well-run program. One with 40 findings where 35 are still open and nobody has decided anything is not, even though the second has fewer formal acceptances. Those accepted findings are also audit evidence, which our guide to evidence for an OT cybersecurity audit covers.
Why Choose Arista Cyber
We run the conversion workshop as part of our assessments rather than delivering a report and leaving, because a report that does not become a plan has not delivered what the client bought. Our consultants work on live industrial assets and hold functional safety credentials alongside OT security certification, so remediation sequencing accounts for outage windows and safety review rather than assuming either. See our OT risk assessments and our case studies.
Next Steps
If you have an assessment report that has not moved, run the bucketing exercise on it. An afternoon with the right people converts a risk-ordered list into a work plan, and the documentation bucket alone usually produces visible progress inside a month.
Common Questions
Why do OT security assessments fail to produce change?
Because findings are organized by risk and work is organized by budget cycle, outage window and available capacity. Nobody converts between the two, so an accurate report sits unexecuted.
How should we prioritize assessment findings?
By what it takes to fix them, not only by risk rating. Sort into documentation, configuration, outage-dependent and capital. The first two are achievable in 90 days; the others need a window or a budget cycle.
What should we tackle first?
The documentation bucket. It needs no outage, no budget and no procurement, closes a disproportionate share of audit findings, and delivers visible completion early, which is what makes the capital request credible later.
What makes an action likely to be completed?
A named person, a date, and a written acceptance criterion. Actions assigned to a department, or with no definition of done, stay open indefinitely.
Is it acceptable to leave findings unremediated?
Yes, provided the decision is explicit: documented, with reasoning, accepted by someone with authority and given a review date. Formally accepted findings are a sign of a managed program and are useful audit evidence.
|
Have an assessment report that has not moved? Arista Cyber converts assessment findings into an executable plan: bucketing by what it takes to fix, named owners with acceptance criteria, a 90-day sequence and a capital case for what needs one. |
Facing tougher cyber security challenges?
Talk to Arista Cyber about your OT security posture.
Request a consultation