Agency QBRs and client reports
Build the client QBR from reconciled exports, with every claim labelled
ZeroTwo can help turn Google Ads, Meta, HubSpot, GA4 and Search Console exports into a client-facing narrative. What makes it hold up in the room is the discipline around it: agree the metric definitions first, then mark each conclusion as measured, inferred or proposed.
Free plan, no card required. Pro is $29.99/month ($26.99/month billed annually). See pricing. ZeroTwo publishes this page and has not published a worked QBR dataset; the sentence shapes below are placeholders.
- 1. Measured
Read directly from an export, or calculated from one.
Needs: The source file, the filter, the definition and the formula.
[Metric] was [value] against a [target] target.
- 2. Inferred
A reason you propose for something that was measured.
Needs: A second source or a test that supports it. Without one, say it is a hypothesis.
Likely driver: [change in budget, creative or season].
- 3. Proposed
What you recommend next, including any projected impact.
Needs: Stated assumptions and a low, base and high range.
Move [amount] to [channel]; expect [range], assuming [rate].
Define the metric before the model touches it
Cross-channel CAC and ROAS only mean something when the cost scope, the revenue source and the attribution window are agreed. Settle these six with the client, write them on one page, and give that page to ZeroTwo with the exports.
| Metric | Agree this with the client | The choice that changes the number | Where the inputs live |
|---|---|---|---|
| CAC | New customers in the period divided by the acquisition cost you both agree counts. | Media spend only, or media plus agency fees, tooling and creative. New customers only, or returning ones too. | Ad platform spend exports and the CRM closed-won list. |
| ROAS | Attributed revenue divided by ad spend, per platform. | Platform-reported revenue or CRM-matched revenue. Attribution window. View-through conversions in or out. | Ad platform exports and CRM revenue. |
| Pipeline | Value of opportunities created in the period, by source. | Which source field is the truth (first touch, last touch or sales-assigned). Created date or close date. Stage cutoff. | The deals export from HubSpot or your CRM. |
| Attribution window | The days between a touch and a conversion that still counts. | Defaults differ between platforms and analytics tools, so two dashboards can disagree about the same sale. | The settings of each ad platform and analytics tool. |
| Organic cost | What counts as the cost of organic search results. | Calling it a $0 media cost hides content, SEO and development time. Decide whether to allocate that time. | Time tracking or the retainer allocation. |
| Blended figures | Whether blended ROAS divides total revenue by total spend or averages the platform figures. | Conversions counted by more than one platform have to be de-duplicated first. | The reconciliation sheet. |
Five sources, two connectors, three file imports
Be clear with the client about how each source gets in. The connectors page lists what is available. For anything else, export it and attach the file.
Google Ads
Listed on the connectors page. Exports also work.
HubSpot
Listed on the connectors page. Exports also work.
Meta Ads
Not on the connectors list. Bring it in as an exported file.
GA4
Not on the connectors list. Bring it in as an exported file.
Search Console
Not on the connectors list. Bring it in as an exported file.
Five steps, each with something you verify
For each step: what to ask ZeroTwo, and what you check before moving on.
Reconcile the exports
Ask ZeroTwo
Combine these exports into one sheet: one row per channel and month, with spend, platform-reported conversions and CRM-matched conversions, plus the file name and date range each row came from.
You verify
Totals match each platform's own total for the same date range. Write down any gap instead of absorbing it.
Diagnose against targets
Ask ZeroTwo
Compare actuals with the targets in this file and label every finding measured or inferred.
You verify
Every inferred finding has a second source or becomes a stated hypothesis.
Write the narrative
Ask ZeroTwo
Write the executive summary and what-happened sections for a CMO. Keep the labels visible and give no figure without its source file.
You verify
Read every number in the draft against the reconciliation sheet.
Build recommendations with ranges
Ask ZeroTwo
For each recommendation give the description, rationale, low, base and high impact with the assumptions behind each, the budget and an owner.
You verify
Placeholder assumptions are replaced with figures you can source, or the recommendation is reworded.
Review with the client
Ask ZeroTwo
List the decisions needed from this meeting and which findings each one rests on.
You verify
Measured, inferred and proposed are shown separately, and the definitions page was agreed first.
Approved recommendations do not end with the QBR. Take them into agency campaign planning to turn each one into an owned workback.
Show the range behind any projected impact
A recommendation such as adding budget to a channel invites the question "what will it return?" A single projected figure with no working is a claim, not an analysis. Build it as a range from assumptions you can source.
The shape of the calculation
Projected pipeline = added budget ÷ cost per lead × lead-to- opportunity rate × average opportunity value
Run it three times, with low, base and high assumptions, and show all three. This is a template: the cells hold placeholders, not results from a real account.
| Assumption | Low | Base | High | Source |
|---|---|---|---|---|
| Cost per lead at the new spend level | [your figure] | [your figure] | [your figure] | Recent exports for this channel. Costs usually rise as spend scales. |
| Lead-to-opportunity rate | [your figure] | [your figure] | [your figure] | CRM funnel report for the same period. |
| Average opportunity value | [your figure] | [your figure] | [your figure] | CRM deals export. |
| Lag between spend and pipeline | [your figure] | [your figure] | [your figure] | Historical created-date gaps in the CRM. |
| Projected pipeline range | calculated | calculated | calculated | Formula above, shown beside the range in the deck. |
Prompts grouped by the label they produce
Each prompt opens in ZeroTwo with the text filled in. Attach the exports and your definitions sheet first.
Measured
- Cross-channel summaryNormalize these Google Ads, Meta, GA4 and HubSpot exports into one summary of CAC by channel, ROAS, pipeline and trial-to-paid rate. Use the definitions on the attached sheet, and show the formula and rows behind every figure.
- CAC by channelFrom these spend and conversion exports, calculate CAC for each channel using only the cost items on the attached definitions sheet. Show the formula, and note where average CAC may not hold as spend scales.
- Monthly client reportWrite a 1-page monthly report with 3 KPIs against target, 2 wins, 1 concern and next month's priorities, in plain English. Every KPI cites its source file.
Inferred
- Performance versus targetsDiagnose Q1 performance against the targets in this file: where we over- and under-performed, with the likely driver for each. Label each driver inferred and say what second source would confirm it.
- Attribution gapsFrom this attribution and conversion-path data, explain which channels drove pipeline and where the attribution model may undercount organic or email. State the attribution window you assumed.
- SEO contribution estimateEstimate the pipeline contribution from organic search using traffic, conversion rate and deal value. Present it as an estimate with its assumptions, and keep it separate from the paid channel figures.
Proposed
- QBR narrativeWrite the QBR narrative with an executive summary, diagnosis, learnings and Q2 recommendations. Keep the measured, inferred and proposed labels on every conclusion.
- Recommendations with rangesWrite 4 Q2 recommendations with rationale, a low, base and high impact range with its assumptions, budget and owner for each.
- Q2 risksIdentify the 3 biggest risks for this client in Q2 and draft a mitigation plan for each, naming what would tell us the risk is materializing.
Before the meeting
A QBR is a diagnosis with a plan attached, not a data dump. Run this list on the draft before it reaches the client.
- Every figure traces to a named export and date range.
- The client agreed the metric definitions before seeing any result.
- The attribution window is printed next to every ROAS and pipeline figure.
- Each conclusion is labelled measured, inferred or proposed.
- Projections appear as a range, with the assumptions beside it.
- Anything the data does not support is removed or marked as a hypothesis.
What a reconciliation needs from your plan
The practical limit is files per task. See the pricing page for the full comparison.
Free
$0/mo
Fine for a one-file monthly summary.
- 100 free credits to start
- 1 file upload per message
Plus
$14.99/mo
For reconciling several exports in one task.
- Multiple file uploads per task
- Memory unlocked
Pro
$29.99/mo
$26.99/mo billed annually
For an agency reporting on many clients.
- Unlimited projects
- Expanded memory
Questions before you build a QBR with AI
Does ZeroTwo connect directly to GA4, Meta and Search Console?
Google Ads and HubSpot appear on the connectors page. GA4, Meta Ads and Search Console do not, so bring those in as exported files and check the totals against each platform before you rely on them.
How do I stop the draft from containing numbers I cannot defend?
Ask for the formula and the rows behind each figure, then spot-check them against the source export. Reject any number that has no source, and keep the measured, inferred and proposed labels on the page the client reads.
Can it estimate the pipeline impact of a budget change?
It can calculate one from assumptions you supply, and you should present it as a low, base and high range, not a single figure. It is a projection that depends on the cost-per-lead, conversion and deal-value assumptions, not a promise.
Which job does this page cover?
The client-facing performance narrative and the decision meeting that follows it. Recurring metric assembly is a different job; the scheduled tasks page covers running work on a timetable. Execution planning for the next quarter is covered on the campaign planning page.
Which plan do I need?
Free allows one file upload per message, which is tight when a reconciliation needs several exports. Plus lists multiple file uploads per task, and Pro lists unlimited projects for an agency with many clients.
Around the QBR
This page covers the client-facing narrative and the decision meeting. Neighbouring jobs have their own pages.
- Agency campaign planning
Turn the recommendations the client approves into an owned workback, a channel plan and a kickoff pack.
- Agency account audits
A structured baseline of a client's accounts, which gives the first QBR something to be measured against.
- Agency social listening
Add competitor and brand-mention context alongside the channel results.
- AI for marketers
For individual marketing tasks outside agency client reporting.
Take last quarter's exports and label every conclusion
Start with one client, write the definitions page, and see which claims in your current QBR survive the labels.