A 30-person AI agency in San Francisco had a beautiful dashboard. It showed revenue, employee count, client logos, and project status in a slick interface that cost $15,000 to build. The problem was that nobody looked at it. It hung on a screen in the office and was shared in a monthly email that most people deleted. When the agency started losing money on projects, the dashboard did not catch it. When a key engineer was overloaded and burning out, the dashboard did not flag it. When the sales pipeline dried up, the dashboard showed healthy revenue from existing contracts but gave no warning about what was coming.
They rebuilt the dashboard from scratch, this time starting with a different question. Instead of "what metrics should we display?" they asked "what decisions do we make regularly, and what data would improve those decisions?" The new dashboard had fewer metrics, uglier charts, and cost nothing to build because it ran on a simple spreadsheet. But every metric on it drove a specific action, and the leadership team reviewed it every Monday morning.
Within two quarters, project margins improved by 15 percentage points, team utilization stabilized, and the agency caught a pipeline gap three months before it would have become a cash flow crisis. The dashboard did not make these improvements happen. The decisions it informed did.
Why Most Agency Dashboards Fail
Agency dashboards fail for three predictable reasons, and recognizing them saves you from building another wall decoration.
Vanity metrics. Total revenue, client count, and employee headcount feel good to look at but rarely inform specific decisions. Knowing your revenue is $2.4 million does not tell you whether you should hire, raise prices, or change your service mix. These metrics are outcomes that tell you where you have been, not leading indicators that tell you where you are going.
Too many metrics. A dashboard with 40 metrics is not a dashboard. It is a data dump. The human brain can focus on 5-7 things at once. If your dashboard has more than 12-15 metrics, nobody is processing all of them, and the most important signals get lost in the noise.
No connection to action. The ultimate test of a metric is: "If this number changes, what would we do differently?" If the answer is nothing, the metric does not belong on your dashboard. Every metric should have a clear trigger: when it crosses a threshold, someone takes a specific action.
The Four-Quadrant Dashboard Framework
Organize your KPI dashboard into four quadrants, each representing a critical dimension of agency health. Together, they give you a complete picture without overwhelming you.
Quadrant 1: Financial Health
These metrics tell you whether your business model is working.
Monthly Recurring Revenue (MRR) or Contracted Revenue.
- What it is: Revenue from signed contracts for the current month
- Why it matters: Shows your baseline financial position
- Target: Varies by agency, but should cover at least 80% of fixed costs
- Action trigger: If MRR drops below 70% of fixed costs, accelerate sales activities immediately
Gross Margin by Project.
- What it is: (Project revenue minus direct labor cost) divided by project revenue, for each active project
- Why it matters: Identifies profitable and unprofitable projects in real time, not at the end
- Target: 45-60% for AI agency projects
- Action trigger: Any project below 30% needs immediate investigation. What is driving the overrun?
Revenue per Employee.
- What it is: Total monthly revenue divided by total headcount (including contractors)
- Why it matters: Measures team productivity and pricing effectiveness
- Target: $15,000-$25,000 per person per month for AI agencies
- Action trigger: Below $12,000 suggests overstaffing or underpricing. Above $28,000 suggests the team is overloaded.
Accounts Receivable Aging.
- What it is: How long invoices have been outstanding, bucketed by 30/60/90+ days
- Why it matters: Cash flow problems usually start as AR problems
- Action trigger: Any invoice over 60 days gets an escalation call. Total AR over 90 days exceeding 10% of monthly revenue triggers a collections review.
Quadrant 2: Delivery Performance
These metrics tell you whether you are delivering quality work on time.
On-Time Milestone Delivery Rate.
- What it is: Percentage of project milestones delivered by the committed date
- Why it matters: Late milestones compound into late projects, which erode client trust and margins
- Target: 85-95%
- Action trigger: Below 80% indicates a systemic problem: either you are overcommitting, under-resourcing, or poorly estimating. Below 70% is a crisis.
Project Health Score.
- What it is: A composite score for each project based on schedule status, budget status, client satisfaction, and team confidence
- Why it matters: Provides an early warning system for troubled projects
- Calculation: Rate each dimension as Green (on track), Yellow (at risk), or Red (off track). Any project with two or more Yellows or any Red needs attention.
- Action trigger: Red projects get a recovery plan within one week. Yellow projects get a check-in within the week.
Client Satisfaction Score.
- What it is: A simple rating from each client, collected monthly or at milestone completion
- Why it matters: Satisfied clients renew, refer, and expand. Unsatisfied clients churn.
- Target: 8.5+ out of 10 average
- Action trigger: Any client below 7 gets an immediate outreach from the project lead. Average below 8 triggers a delivery process review.
Rework Percentage.
- What it is: Hours spent on rework divided by total project hours
- Why it matters: Rework is the most expensive type of work and usually indicates process or communication failures
- Target: Below 10%
- Action trigger: Above 15% triggers an investigation into root causes. Is it requirements clarity? Code quality? Testing gaps?
Quadrant 3: Team Health
These metrics tell you whether your people are productive, engaged, and sustainable.
Team Utilization Rate.
- What it is: Billable hours divided by available hours, averaged across the team
- Why it matters: The leading indicator of both profitability and burnout
- Target: 60-70% agency-wide average
- Action trigger: Below 55% means you need more work or fewer people. Above 80% means people are burning out and you need to hire.
Individual Utilization Extremes.
- What it is: The highest and lowest utilization rates on the team
- Why it matters: Averages hide problems. One person at 95% utilization is in trouble even if the average is 65%.
- Action trigger: Anyone above 85% for two consecutive weeks needs workload redistribution. Anyone below 40% for two consecutive weeks needs a conversation about role and projects.
Open Positions and Days to Fill.
- What it is: Number of open roles and how long each has been open
- Why it matters: Unfilled positions create overload on the existing team
- Target: Less than 45 days to fill
- Action trigger: Any position open longer than 60 days needs a hiring process review. Are the requirements too narrow? Is the compensation competitive?
Voluntary Turnover Rate (Trailing 12 Months).
- What it is: Number of voluntary departures divided by average headcount over the past 12 months
- Why it matters: Losing people is expensive and disruptive. Patterns in turnover reveal systemic issues.
- Target: Below 15% annually
- Action trigger: Above 20% triggers a retention investigation. Exit interview data should be reviewed for patterns.
Quadrant 4: Pipeline and Growth
These metrics tell you what is coming, not just what is happening now.
Weighted Pipeline Value.
- What it is: Sum of all open deals multiplied by their probability of closing (based on stage)
- Why it matters: Predicts future revenue and capacity needs
- Calculation: Use historical close rates by stage. If 40% of deals in your proposal stage close, a $100K deal at that stage is worth $40K in weighted pipeline.
- Action trigger: If weighted pipeline drops below 2x your quarterly revenue target, increase sales activity.
Pipeline Velocity.
- What it is: Number of qualified deals multiplied by average deal size multiplied by win rate, divided by average sales cycle length
- Why it matters: Shows whether your pipeline is speeding up or slowing down
- Action trigger: Declining velocity for two consecutive months warrants investigation. Is deal size shrinking? Is the cycle getting longer? Is the win rate dropping?
Lead-to-Qualified Ratio.
- What it is: Percentage of new leads that become qualified opportunities
- Why it matters: Measures the quality of your lead generation and initial qualification
- Target: 20-30% for AI agencies
- Action trigger: Below 15% suggests you are attracting the wrong leads. Above 40% suggests you might be qualifying too loosely.
Client Expansion Revenue.
- What it is: Additional revenue from existing clients (new projects, expanded scope, renewals at higher rates)
- Why it matters: Expansion revenue is cheaper and faster to close than new client revenue
- Target: 30-40% of total revenue should come from existing clients
- Action trigger: Below 20% suggests you are not cultivating existing relationships. Invest in account management.
Building the Dashboard: Practical Implementation
Data Sources
For each metric, identify where the data comes from:
- Financial metrics: Accounting software (QuickBooks, Xero) or billing platform
- Delivery metrics: Project management tool (Asana, Linear, Jira) and time tracking tool
- Team metrics: Time tracking tool, HR system, and manual tracking (for satisfaction scores)
- Pipeline metrics: CRM (HubSpot, Pipedrive)
Tool Options
The Spreadsheet Dashboard (Best for agencies under 15 people). A well-designed Google Sheet or Excel workbook with four tabs (one per quadrant) and a summary page. Update it weekly. It costs nothing, requires no technical setup, and is infinitely customizable. Do not underestimate the spreadsheet.
Databox or Geckoboard (Best for agencies 15-30 people). These tools pull data from your existing SaaS tools and display metrics in a configurable dashboard. They look polished and update automatically. Budget $100-$300 per month.
Looker Studio or Metabase (Best for data-savvy agencies). If you have a data warehouse or strong SQL skills, these tools let you build custom dashboards from raw data. More powerful but requires more setup and maintenance.
Custom dashboard (Best avoided). Building a custom dashboard is a project that distracts from client work and is never quite finished. Use an existing tool unless you have a very specific need that nothing else addresses.
Update Cadence
- Financial metrics: Updated monthly (after books close)
- Delivery metrics: Updated weekly
- Team metrics: Updated biweekly or monthly
- Pipeline metrics: Updated weekly
Review Cadence
Weekly leadership review (30 minutes). Review Quadrant 2 (delivery) and Quadrant 4 (pipeline) every Monday. These are the fast-moving metrics that require weekly attention.
Monthly business review (60 minutes). Review all four quadrants in depth. Look at trends over the past three months. Identify metrics that are trending in the wrong direction and assign action items.
Quarterly strategic review (2-3 hours). Use the dashboard data to inform quarterly planning. Which metrics need to improve? What initiatives will drive those improvements? Are the dashboard metrics still the right ones, or has the business changed?
Setting Targets and Thresholds
Every metric needs three values: a target, a warning threshold, and a critical threshold.
Target: The number you are aiming for. This is your definition of "good."
Warning threshold: The number that triggers attention. Something is moving in the wrong direction and needs investigation.
Critical threshold: The number that triggers immediate action. Something is broken and needs fixing now.
Example for utilization rate:
- Target: 65%
- Warning: Below 55% or above 75%
- Critical: Below 45% or above 85%
Color-code your dashboard accordingly: green for on-target, yellow for warning, red for critical. This visual system lets you scan the entire dashboard in 30 seconds and immediately see what needs attention.
Evolving Your Dashboard Over Time
Your dashboard should not be static. As your agency grows and your priorities shift, the metrics should evolve.
Quarterly review of metrics. Ask three questions about each metric:
- Did this metric influence any decisions this quarter?
- Is it still measuring the right thing?
- Should it be replaced with something more relevant?
Adding metrics. When you identify a new decision that needs data, add a metric. But also remove one. The dashboard should not grow beyond 15 metrics.
Removing metrics. If a metric has been green for four consecutive quarters, it might not need dashboard-level visibility anymore. Move it to a secondary report and replace it with something more actionable.
Adjusting targets. As your agency matures, targets should increase. If you consistently hit 90% on-time delivery, raise the target to 95%. Targets that are always met do not drive improvement.
Your Next Step
This week, create a single spreadsheet with one row per metric described above. For each metric, fill in three columns: the current value (your best estimate), the target value, and where the data comes from. You will quickly discover which metrics you can already track with existing data and which require new data collection. Start your dashboard with the metrics you can populate today, even if it is only half of them. A dashboard with eight real metrics is infinitely better than a plan for sixteen metrics that never gets built. Review it next Monday morning with your leadership team and commit to weekly reviews going forward.