The peer decision path

What it is

Every Decision Verinode surfaces is a recommendation built from your own data. The peer decision path answers the next question you would ask a fractional COO: has anyone else been here, and did it work out? It is a compact stat block that sits underneath a decision and reports, for operators in a comparable cohort, how many faced the same signal recently, how many acted on it, and of those, how many saw the situation improve afterward.

Verinode never tells you what to do here. This block only reports what the network already chose and what happened next. You still decide whether to act, park, or dismiss the decision in front of you.

The block is powered by intelligence.peer_decision_paths, a nightly rollup that never stores an operator ID, an operator hash, or any raw peer identity, only counts. Consent matters too: an operator's activity only rolls into a cohort's counts if that operator has consented to benchmark contribution. If you have not opted into benchmark contribution yourself, your own activity is never counted in any cohort either.

Where to find it

The peer decision path appears in two places, both driven by the same underlying data:

• On the Decision card itself. Open Decisions from the sidebar (iq.verinode.ai/decisions), or scroll the Feed, where the same cards surface inline. A compact, three-line version of the block sits below the card's recommendation, separated by a hairline rule. This is the "card" variant.
• Inside the decision workspace. Click into any decision to open its full workspace at /decisions/[id]. There, the block expands into a larger panel headed What peers did, positioned directly above the Related Research section, because the two answer adjacent questions: what did peers do, and what does the research backing this action say. This is the "panel" variant.

If a decision has no peer coverage yet, in the sense that no cohort has accumulated enough activity on that exact signal, the block does not render at all, on the card or in the workspace. There is no placeholder, no "not enough data" message, and no empty box. The section simply is not there. This is deliberate: Verinode never surfaces internal thresholds or "insufficient data" language to operators.

How your cohort is chosen

The peer decision path always compares you to a cohort, never to the whole network undifferentiated, and never to a single named peer. Verinode builds your cohort from two facts on your operator profile: your state and your size band (solo, small, mid-size, or large, based on employee count). It then looks for the most specific match available, in this order:

1. State and size together ("operators like you in [your state]"). Verinode deliberately does not name the size band in this label, even though it is matching on it, because spelling out both dimensions at once starts to feel like it is describing you specifically rather than a group.
2. State alone ("operators in [your state]").
3. Size band alone ("small operators nationally," "mid-size operators nationally," and so on).
4. National ("operators nationally"), the broadest fallback, always available once any network-wide rollup exists.

Verinode always uses the narrowest cohort that has enough peer activity to report safely. If your state has too little recent activity on a given signal to form its own cohort, the block falls back to the next broader scope automatically. You never see which scope was used, and you never see how many peers sit behind it beyond the headline count on the card. The label itself is the only clue, and it is intentionally general.

The card variant

On the Decision card, the block reads as three short lines of text under a hairline separator:

1. "[N] [cohort label] faced this in the last 90 days." For example, "14 operators like you in TX faced this in the last 90 days." The bolded number is the cohort size, how many peer operators encountered this exact signal type in the trailing 90-day window. The cohort label is one of the four forms described above.
2. "[N] acted ([X]%)." How many of those peer operators took action on the signal, and what share of the cohort that represents. If a verified outcome exists for at least one of those acted operators, the line continues: "Of those, [N] saw improvement ([X]%)." That second count and percentage are scoped to the operators who acted, not the whole cohort, so 5 improved out of 8 acted reads as 63%, not a fraction of the 14 who faced the signal.
3. A modal-plan line, when one exists. If a clear majority or plurality of the acted operators followed a similar shape of plan, a line appears summarizing it in plain language, either "Most chose a [plan shape]." when the share is decisive, or "The most common path was a [plan shape]." when it is a plurality rather than a clear majority. Plan shapes are described qualitatively, things like "single-message outreach," "single-doc plan," "phased rollout over a month," never a literal list of every peer's individual steps. This line is omitted entirely when no plan shape stands out clearly enough to be worth naming.

If no acted operator yet has a verified outcome, the "Of those, X saw improvement" clause is dropped from line 2, and the card simply stops after the acted count and percentage.

The panel variant (workspace)

Inside the decision workspace, the same underlying numbers render as a fuller panel:

• A header row reads What peers did, with last 90 days beside it, naming the trailing window the whole panel covers.
• Below that, a sentence: "[N] [cohort label] faced this signal."
• A three-column stat row follows:

- Acted, a large number with "acted ([X]%)" beneath it, the count and share of the cohort that took action. - Improved, in green, with "improved ([X]%)" beneath it, when outcome data exists. This is the count and share, among acted operators, whose situation moved in the right direction afterward. - No change or worse, a combined count of the acted operators whose outcome was flat or regressed, so the three numbers together (improved, no change or worse, and any still-pending) account for everyone who acted.

• When no acted operator has a verified outcome yet, the two outcome columns collapse into a single wide note instead: "Outcomes for the [N] acted operators are still resolving. Verified results land within 60 days of action." That 60-day figure is the verification window Verinode holds open before an acted-on decision's outcome is considered settled enough to count.
• The modal-plan line, when one exists, repeats below the stat row in the same wording as the card variant.
• A partial-coverage note, when some but not all acted operators have a verified outcome: "Outcome data covers [N] of [M] acted operators, the rest are still in the verification window." This tells you the improved/no-change-or-worse split you are looking at is based on a subset of everyone who acted, not the full acted count, because some outcomes simply have not had time to resolve yet.

Reading the numbers together

The three layers, faced, acted, improved, are a funnel, and each step tells you something different:

• A high faced count with a low acted share tells you most peers are sitting on this kind of signal rather than moving on it, which is worth knowing before you assume urgency.
• A high acted share with a strong improved share is the closest thing to a track record Verinode can show you: peers who saw this, moved, and it worked.
• A high acted share with a weak improved share, or a large "no change or worse" bucket, is a signal to look harder at your own plan before copying the network's default move, not a reason to skip acting altogether.
• The modal-plan line tells you what shape of response was most common, not whether it was the best one. A plan can be popular and still not be the right fit for your situation.

None of this replaces your own judgment. Verinode reports what happened across the cohort; you are the one who knows your crew, your cash position, and your carrier relationships well enough to decide whether the peer path applies to you.

1. 1Open a Decision card in the Feed or Decisions list and look under the hairline rule for the peer stat lines.
2. 2Note the cohort label to understand who you are being compared to (your state, your size band, or the national pool).
3. 3Read the acted percentage first: a low share means most peers have not moved yet.
4. 4If an improved percentage is showing, weigh it against the "no change or worse" side before assuming the peer path is a strong endorsement.
5. 5Click into the decision to open the workspace for the fuller What peers did panel, including the partial-coverage note if outcomes are still resolving.
6. 6Use the modal-plan line, if present, as a starting shape for your own plan, then adjust it in the workspace's plan builder for your situation.

Empty state

There is no visible empty state. When Verinode has no cohort coverage yet for a signal's exact domain and type, the peer decision path block does not render, on the card or in the workspace. No placeholder text, no grayed-out box, no "insufficient data" message appears. The rest of the decision, the recommendation, the plan, the impact estimate, renders normally with the peer section simply absent.

Data sources

Related reading

• The decision workspace
• Acting on decisions
• The Feed
• Benchmarks overview
• How benchmarks work
• Reading a benchmark