"Adherence: how often your jobs actually follow the SOP"

What it is

Writing an SOP is the easy part. The hard part is knowing whether the field is actually running it. The Adherence panel answers that one question directly: of the jobs your team closed out in this category over the last 90 days, how many followed the documented SOP step-for-step, and how many drifted from it?

Verinode does not watch your crews or grade anyone personally. It reads the job executions that flow in through your closeout webhook, checks each one against the SOP's steps, and counts what it finds: a job with no recorded deviation counted as followed, a job with one or more recorded deviations counted as not followed. The panel turns that count into a number, a trend, and a plain-language read. You decide what to do with it, whether that is retraining a step, simplifying the SOP, or leaving it alone because the SOP is fine and the field just needs a refresher.

Where to find it

Open Processes from the sidebar (Operations section), at /processes. Click any tile in My SOPs to open its detail card. The Adherence panel is its own section, titled Adherence (90 days), sitting below the Status section (draft or adopted-from-standard framing, when either applies) and above Credentials this SOP implies.

The panel is scoped to one SOP at a time. It only counts jobs in that SOP's category that closed out and linked back to it, so a mitigation SOP's adherence number has nothing to do with your fire or reconstruction SOPs.

The headline count

At the top of the panel, a large number reads followed/total, for example 18/24. Read literally: of the last 24 jobs in this category that closed out and linked to this SOP in the last 90 days, 18 ran with no recorded deviation from the documented steps.

Beside the count, a percentage repeats the same ratio (18/24 renders as 75%). The percentage is colored to make the overall picture readable at a glance without doing the math:

• Green, 80% or higher. Most jobs are running the SOP as written.
• Amber, 50% to 79%. Adherence is mixed.
• Red, under 50%. More than half the jobs in the window deviated from the documented steps.

These are the same three bands the sparkline dots and the closing guidance sentence use, so the color language is consistent from the top of the panel to the bottom.

The cohort confidence label

Next to the headline count, a pill names how much weight to put on that number: Early signal, Indicative, Observed, or Verified. This is the same labels-over-gating ladder used elsewhere on the platform: rather than hiding the ratio until enough jobs have closed out, Verinode always shows the number and instead tells you, in plain language, how much confidence to place in it.

Read the ladder as a progression from thin to thick data:

• Early signal means only a handful of jobs have closed out and linked to this SOP so far. The ratio is real, but a single unusual job can swing it a long way. Treat it as a first read, not a verdict.
• Indicative means enough jobs have come through that a pattern is starting to take shape, though it can still move with the next few closeouts.
• Observed means the sample has grown enough that the ratio is holding up as more jobs land, closer to a settled read on how this SOP actually runs in the field.
• Verified means the widest sample this ladder tracks: the ratio has been tested against enough closed-out jobs that it reads as a dependable account of how this SOP performs, not a snapshot that a few outlier jobs could tip over.

The label strengthens automatically as more jobs close out against this SOP. There is nothing to configure and nothing to unlock. It moves on its own as your execution data accumulates.

The 13-week sparkline

Below the headline count, a small chart plots adherence week by week across the same 90-day window, split into 13 Monday-to-Sunday weeks.

Each week is one dot on the chart:

• Height encodes that week's adherence ratio. A dot near the top of the chart means most jobs that closed out that week followed the SOP; a dot near the bottom means most deviated.
• Color follows the same green (80%+), amber (50-79%), red (under 50%) bands as the headline percentage.
• A thin stem connects each dot down to the baseline, so you can read the shape of the trend at a glance, rising, falling, or flat, rather than parsing tooltips one at a time.

Hover any dot to see the exact numbers for that week: the date the week starts, the followed/total count, and the percentage.

Missing-data weeks. Not every week in the window necessarily has a job that closed out against this SOP. A week with zero closeouts renders as a small, faint dot sitting on the baseline, visually distinct from a week that scored low. This distinction matters: a faint baseline dot means "no signal that week," not "adherence collapsed that week." Reading the sparkline correctly means checking dot color and height together with dot faintness. A run of faint dots tells you the category is not producing enough closeout volume to trust the trend yet, which is a different problem from a run of red dots, which tells you the field is deviating from the SOP.

The variance line

When the panel has enough data, a line beneath the sparkline reads:

Closeout-webhook jobs ran [on plan / N% over plan / N% under plan] on actual vs estimated minutes.

This measures something different from the followed/total count above it. Adherence asks whether the crew ran the documented steps. Variance asks whether the job took as long as the SOP's steps estimated it would, averaged across every closeout-webhook job in the window that has both an estimated and an actual minutes figure recorded.

• On plan appears when the average is within about 5% of the estimate either way, close enough to call it accurate.
• N% over plan means jobs are, on average, taking longer than the SOP's steps estimate.
• N% under plan means jobs are wrapping faster than estimated.

Because this line depends on jobs carrying both an estimated and an actual minutes figure, it does not appear at all when no closeout-webhook job in the window has that data recorded yet. There is no separate empty-state message for it, the line is simply omitted.

The closing guidance line

The panel ends with one sentence that changes based on the headline ratio. It is Verinode's read on what the number means, not an instruction:

• Under 50%: "Crews are deviating from the documented SOP on more than half the jobs in this window. Walk a few jobs end-to-end before second-guessing the SOP, adoption-without-adherence is more common than the SOP being wrong."
• 50% to 79%: "Adherence is on the way up but not consistent yet. The deviation captures on the Jobs tab tell you which steps the crew skips most."
• 80% and above: "Field crews are running the SOP step-for-step. The drift on the sparkline points to weeks worth investigating, not the average."

The pattern across all three bands is the same: a low ratio is framed first as an adoption question (is the field actually running this SOP at all) before it is framed as a quality question (is the SOP itself wrong). That ordering matters. Rewriting a perfectly good SOP because the field never adopted it wastes effort on the wrong fix.

Empty state

Before any job has closed out and linked to this SOP, the panel does not show a zero count or an empty chart. It reads, verbatim:

"No execution data yet. When jobs in this category close out and link to this SOP, adherence + drift signal land here."

If the SOP has no category recorded (uncommon, but possible on a hand-built SOP that has not had a category assigned), the second sentence reads instead:

"When jobs close out and link to this SOP, the adherence rate lands here."

Both versions mean the same thing: the panel is not broken, it is waiting on closeout data. Nothing to configure here either, once jobs in the category start closing out through the closeout webhook and linking to this SOP, the panel switches from the empty state to the live count on its own.

How to use it

Start at the headline count and color, then work down:

1. 1Read the followed/total count and its color. This is your fastest read on whether this SOP's field execution needs attention.
2. 2Check the cohort label before treating the ratio as settled. An Early signal or Indicative ratio is a first read, not a verdict.
3. 3Scan the sparkline for faint (no-data) weeks versus colored (scored) weeks. A trend built on a handful of scored weeks with several faint gaps is thinner evidence than a trend where every week has a dot.
4. 4Look at the shape of the sparkline, not just this week's dot. A SOP climbing from red to green over 13 weeks is a different story than one holding steady at green, or one that just dropped after months of green.
5. 5Read the variance line, when present, alongside adherence. A SOP with high adherence but jobs running well over the estimated minutes suggests the steps are being followed but the time estimates on those steps need revisiting.
6. 6Act on the closing guidance line's framing: below 50%, check adoption first; between 50% and 80%, use the deviation captures on the Jobs tab to find the specific steps crews skip; above 80%, treat the sparkline's dips as the interesting cases, not the headline average.

Related reading

• The SOP detail card
• Processes overview
• Connecting your data
• How benchmarks work
• Benchmarks overview
• Understanding your margin