Read the Suggested Plan

Open the Suggested Plan

1. 1

   Open Insights in the left-hand menu, then Suggested Plan.

   The other views are Executive Read, Data Audit, Analysis and Field Intelligence. Report Templates stays on the tool’s own page under Data sources.

2. 2

   Pick the tool you want the plan for along the top.

3. 3

   Set the Created filter to the period you plan against.

4. 4

   Wait for the plan to draw.

   It reads your audit, your submissions and your profile, and it waits up to 40 seconds for the Analysis suites. An analysis slower than that contributes nothing, and What this plan could draw on says so in as many words.

5. 5

   Click Rebuild after you change anything.

   The plan is held for five minutes per tool and per date range, so re-opening the tab is instant. Rebuild is the only thing that ignores that. The time underneath is when the plan you are looking at was built.

The headline, the brief and the line between them

The top card carries four things. Read them in order.

• A badge reading Suggested plan or Readiness plan. The second means MarqIQ refused to rank anything — see When you get a readiness plan instead.
• The headline: the title of your first investment bet, or of your first fix when there is no bet, or “Where to spend the next quarter” when neither fired.
• One line saying what the plan covers — for example “Read all 77,662 people.” or “Read 12,400 people created in the last 12 months (of 77,662 in all).” That is the population every figure below rests on, apart from the two tiles named above.
• A short brief. It appears a beat after the numbers and never holds them up.

The line under the brief says where the brief came from. Written by AI from this plan — every number checked against it means a model wrote it. Summary from the plan means MarqIQ’s own wording shipped instead.

Where the numbers come from

This is the short answer to “can I trust it?”. Four rules hold across the whole tab.

• No model chooses, computes or words any figure, verdict or ordering. The plan is worked out the same way every time from the same inputs. AI writes one paragraph, and only over numbers the plan already contains.
• Every figure names what it was measured against. A card’s evidence rows read “value of denominator”: a rate always arrives with the number of records it was computed on, never on its own.
• There is no benchmark of any kind. Every comparison is against your own overall rate, never against an industry average, because MarqIQ holds no cross-customer dataset.
• A rate that cannot clear a volume floor and a confidence bound is reported as not enough to go on, never quietly dropped. A missing section reads as “we checked and it was fine”, which is worse than an absent one.

The five tiles

The strip across the top frames everything under it. The three columns below are what each tile counts and, in the third, the population it counts it on — the part that decides whether two tiles can be put in the same sentence.

Colour is set on stated thresholds, not on judgement. Lead source fill is green at 70% or more, amber from 50% to 69%, red below 50%. Submissions, no first touch is green under 15%, amber under 25%, red at 25% or more. AI answers that mention you is green at 40% or more, amber from 20%, red below.

The scorecard

Five areas, one grade and one line each: Data foundation, Lead scoring, Lifecycle & handoff, Campaign mix, and Content & enablement. It exists so you can see the shape of the problem before opening a single card.

The line beside the grade is the title of the leading finding in that area, with “(+2 more)” when there are others. Findings that were real but held back by the three-item cap still count towards the grade, so a fourth problem never turns an area into an A.

Where to invest

At most three growth bets, ranked on their own number rather than on severity, and read in the same sitting as the fixes rather than after them. Two different kinds of item compete for those three slots.

• Ranked bets, which compare a segment against your own baseline. There are four of them, and each writes its own headline from your records, so the card is titled with the finding rather than with a fixed name — “Put more into Paid Search — it converts at 2.4× your baseline”, for instance. You see their fixed names, Where to invest next, Your best buyer, from the data, What to stop feeding and Revenue concentration, only under What we could not tell you, where a bet that did not run is listed. All four need enough decided outcomes to rank anything.
• Growth motions, which rank no channel and so need no conversion statistics at all: a win-back to the people you can still email and have stopped talking to, a second way in when one form carries the intake, one campaign tagged properly so a single cohort is clean from the start, and a campaign that creates the behaviour a score will read when scoring is not set up.

Each motion carries its own volume floor, stated on the card: 100 or more quiet and never-engaged people; 30 or more people a form brought in, with one form holding at least half of them; 200 or more people with at least a quarter of them carrying no source; and, where MarqIQ has checked that your tool supports lead scoring and found none set up, 100 or more people.

Where to invest next has to clear three tests before it appears: at least 30 records in the segment, at least 2% of the volume in that field, and a conservative lift of 1.5× or more over your baseline. The lift quoted is the Wilson lower bound, not the raw ratio — the number MarqIQ is willing to defend rather than the flattering one. Your best buyer, from the data puts every firmographic field through those same three tests and combines the winners.

The other two bets are tested differently, and the difference matters when you quote them. What to stop feeding needs 60 or more records on the value, at least 10% of the volume in that field, and a conservative lift of 0.6× or below — the Wilson upper bound this time, so the segment is under baseline even read generously. Revenue concentration runs no rate test at all: it needs two or more values carrying a recorded amount, with the largest holding 40% or more of that recorded revenue.

Open Where to invest next and you get its evidence rows, each with its denominator: the segment’s rate of N records, your overall rate of N records, the conservative lift against baseline, customers above baseline as “produced versus expected”, and the segment’s share of volume. A revenue row appears only where your CRM carried an amount on those records. The other three bets carry their own rows, each written the same way.

What to fix

At most three repairs, in order. Three things decide that order and nothing is re-judged here: severity first, then reach — how much of the analysed population it touches — then effort, so a tie goes to the cheaper action.

Every card folds open on its headline, its owner role, one number and its definition of done. Open it for the full evidence. Inside, each card reads in the same four parts: What we found, What to do, Why it matters, Done when.

• The owner is a role, never a person: Marketing Ops, Demand Gen, Content, RevOps, Sales leadership, CRM admin or Leadership. A recommendation with nobody attached is the one that gets nodded at and never scheduled.
• Done when is a condition the same analysis can confirm on its next run — “Lead source populated on at least 70% of records, with one system as the writer” — never an invented percentage lift. MarqIQ has no counterfactual, so it does not quote one.
• A finding raised by an Analysis suite carries a line reading Found by the … analysis with a See the evidence link straight to the analysis that holds it.

Two more repairs sit under Worth doing alongside, inside Show the working. They are the cheap end of what is left, picked because they can be absorbed without displacing the three above.

This week

The low-effort items on a 30-day horizon, sorted by severity, at most three, each prefixed with the role that owns it. Hand them to whoever has capacity now and ship them while the quarter plan is still being agreed. The section does not appear at all when nothing qualifies.

What sales should know and What RevOps should settle

The same plan restated in two columns, at most four lines each. No claim is added here — both columns are re-cut from findings that already fired.

• What sales should know is a prioritisation change: which segment to work first and the rate that justifies it, the best buyer written from your records, which segment is getting negative scoring weight, and where qualified records are landing with no owner.
• What RevOps should settle is definitions, denominators, handoffs and governance: fix the source denominator before any channel number is quoted, write sourced and influenced into the field descriptions, instrument the handoff with an ownership timestamp and an SLA, agree a survivorship rule before any merge.

The six-month roadmap

Six months as a strip, starting the first of next month. Open a month for its card. The sequence is ordered by dependency, not by your calendar.

Every month card carries an Audience, Channels, Ship (the assets), Measure (the KPI, with its baseline where there is one), a Stop if on anything that is a bet, and Why here. A month that depends on an earlier one says so under After.

The chip beside each theme is the one to read first. From your data means the month rests on something measured. From what you told us means it rests on the business form. Data + what you told us means both. Those deserve different confidence, and hiding the difference is how a roadmap becomes a template with a logo on it.

One roadmap KPI is worth knowing the arithmetic for. The answer-engine month appears only when a usable run shows a mention rate under 25% across at least five prompts, and its target is your own baseline plus exactly 10 points, capped at 100. It states that baseline as a percentage of a named number of answers across a named number of prompts. Where the last run was not usable, the month becomes a re-measure instead, because aiming a quarter at a measurement gap is not a programme.

Above the strip sit two short lines. The first is the premise the roadmap was built on: what you sell and who buys it, where you have told us, and otherwise the sentence “Built from your synced data alone.” The second reads Sharper if you tell us … and names the single most useful thing still missing from the business form.

What this plan could draw on

Twelve rows under Show the working, with a count in the heading reading “n of 12”. Each row is one input the plan could have used, whether it did, the real figure behind it, and — where it did not — the exact setup step that would unlock it.

The unticked rows are the point of the strip. They are the shortlist of what to connect or tell us to make next quarter’s plan sharper, and they are why a plan that lists its own blind spots is the one a sceptical executive can read.

Show the working

One toggle at the foot of the tab, reading Show the working — what this drew on, what else fired, definitions and limits. It holds five things: the coverage strip above, Worth doing alongside, What we could not tell you, Definitions used here and What this plan is not claiming.

Definitions used here

The terms this report uses, as it uses them, so the meeting does not start with an argument about denominators. It always states which of conversion rate or win rate you are reading and what its denominator is. It adds how “sourced” was classified and from which field when the audit could compute that, that stages are a snapshot of where records sit today and not a history of how they moved, and how many records were analysed from which tool over which created window.

What this plan is not claiming

Printed, not footnoted. Every plan carries the no-history limit. A workspace with no opportunity rows also carries a line saying no figure anywhere carries a currency amount. A contact-grain workspace carries a line saying its rates are conversion rates and are never mixed with a deal win rate. A sampled audit says its figures describe the records analysed and not the whole database. Every plan carries the line that there is no cross-customer benchmark anywhere in it.

What we could not tell you

Every check that did not reach the page, with its reason and what would bring it back. Most reasons are written by the check itself and name the figure that failed — “No lead-source field resolved in the analysed records.”, for example. Six reasons are written by the plan rather than by a check, and those six mean different things.

A check suppressed by the outcome gate is listed here too, with the figure that failed the test and the sentence “Sync outcome data and reach at least 30 decided records, then this runs.” A check that threw an error reads “This check could not complete against the current data.” and is retried on the next audit; nothing is dropped in silence.

When you get a readiness plan instead

Before anything else runs, MarqIQ asks whether there is an outcome to measure at all. Any one of four answers stops it ranking:

• Fewer than 500 records were analysed. That is the count under your Created filter, not everything you have synced, so a narrow filter can trip this on its own.
• No outcomes are recorded at all.
• Fewer than 30 records have reached a decided outcome.
• A source field resolved and is populated on under 50% of records, so a channel ranking would describe the recorded minority. A workspace where no source field resolved at all passes this one and is stopped by the checks above instead.

What you get then is a different document, not a shorter one. The badge reads Readiness plan, the headline becomes “Making the next quarter measurable”, and the fix column is titled Make the next quarter measurable with the line “Nothing here ranks a channel; every item makes the next quarter measurable.” Only checks that rank nothing are kept. The growth motions stand beside them, so the quarter is not all repairs.

Work those items first. They are what makes the next plan rankable, and the reason on screen names which of the four stopped it, so you know what to fix.

Which parts stay blank, and for which tools

Nothing here is hidden, but several sections read differently depending on what your tool holds. These are the ones worth knowing before you present the plan.