There is a word that gets used loosely in conversations about PI marketing performance, and the looseness costs firms real money.
The word is “data.”
When a managing partner says “we need better data on our marketing,” they typically mean they need better answers to specific questions: What is our cost per case? Which vendors are performing? Where should we invest more? But what most firms actually build — in response to that request — is more data: more dashboards, more reports, more metrics. More inputs that still do not produce clear answers.
Data and intelligence are not the same thing. Understanding the difference — and actively pursuing intelligence rather than data — is one of the most important operational distinctions in high-performing PI firms.
What Data Is
Data is a recorded observation. “Vendor A sent 180 leads in March.” “Total marketing spend was $340,000.” “43 cases were signed.” “Average response time was 8.4 minutes.”
Each of these statements is accurate. Each is useful at a basic level. None of them, individually, tells you what to do. They are facts in isolation — numbers without sufficient context to drive decisions.
Most PI firms have an abundance of data. Ad platforms produce reports. Vendors send monthly summaries. Case management systems store every lead interaction. Call tracking platforms log every inbound call. The data exists. It is often not collected systematically, it is certainly not connected, but it exists.
What Intelligence Is
Intelligence is what data means when connected, analyzed, and interpreted for a specific decision. “Vendor A sent 180 leads in March. Of those, 9 became signed cases — a 5% conversion rate, which is down from 8.3% in January and 7.1% in February. The cost per case from Vendor A has risen from $3,200 in January to $5,600 in March. Meanwhile, Vendor C is delivering cases at $2,900 with a stable 9% conversion rate. The March data suggests either Vendor A's lead quality is declining or there is an intake handling issue specific to their lead type.”
That is intelligence. It answers specific questions, connects multiple data points, identifies trends, and surfaces a decision — reduce Vendor A's allocation and investigate the conversion decline before the March invoice compounds into April.
The individual data points (180 leads, $340K spend, 43 cases) were present in both the data statement and the intelligence statement. The difference is connection, context, and the presence of a clear decision implication.
| Dimension | Data | Intelligence | |
|---|---|---|---|
| Example | Vendor A sent 180 leads | Vendor A's cost per case rose from $3,200 to $5,600 over 3 months | |
| Provides Context | |||
| Identifies Trends | |||
| Surfaces a Decision | |||
| Connects Multiple Systems |
How the Confusion Costs PI Firms Money
When firms respond to the need for better marketing performance measurement by producing more data, predictable problems follow.
The dashboard proliferation problem
More data often means more dashboards. Marketing gets a dashboard. Intake gets a dashboard. The managing partner gets a monthly report. Each dashboard answers narrow questions about the data set it was built to display. None of them connects the view across data sets. The firm has invested in reporting infrastructure that produces more information without producing more clarity.
The partners still cannot answer “what is our cost per case by vendor?” because that question requires connecting the marketing dashboard to the intake system — and nobody has built that connection. The data exists. The intelligence does not.
The metrics that look good but mean nothing
Data without context often produces optimism that is not warranted. A marketing director reporting 600 leads in March at an average cost of $65 per lead can present data that looks impressive — but if 75% of those leads were immediately disqualified for case type or jurisdiction, the effective cost per workable lead is $260, and the cost per signed case may be catastrophic.
Cost per lead data, in isolation, is not intelligence. It is a single variable that only acquires meaning when connected to conversion data. Firms that optimize for cost per lead — because that is the data they have — often end up with cheap leads that cannot be converted, rather than more expensive leads that produce cases reliably.
The vendor negotiation that goes nowhere
Vendors provide data. That data is accurate in the narrow sense that it correctly reflects what they have measured. But vendor data is not intelligence about vendor performance, because vendors do not have access to your conversion data, your case outcomes, or your settlement results. Vendor-provided reports can tell you what you paid and how many leads arrived. They cannot tell you whether those leads were worth what you paid.
Firms that walk into vendor negotiations with vendor-provided data are negotiating with the vendor's own numbers. That is not a strong position. Firms that walk in with independently calculated cost per case data — their own data, from their own systems — are negotiating with intelligence. The conversations produce very different outcomes.
The Four Questions That Separate Data from Intelligence
A practical way to distinguish data from intelligence in PI marketing is to ask four questions of any metric you are tracking:
1. What decision does this inform? If a metric does not connect to a specific decision — which vendor to fund, which intake rep to coach, which campaign to scale — it is data, not intelligence. Track it if you must, but do not confuse it with an answer.
2. What context does it require? Most metrics only mean something in relation to other metrics. Cost per lead requires conversion rate context. Conversion rate requires lead quality context. Case volume requires case severity context. Intelligence requires connecting the metric to the context that gives it meaning.
3. What is the trend?Point-in-time data describes the present. Trend data enables prediction and early action. A metric that is declining tells you something a static snapshot cannot. When a vendor's conversion rate drops from 9% to 7% to 5% over three months, the trend is the intelligence. The March number alone is just data.
4. What should we do differently because of it?Intelligence produces a clear action implication. If analyzing your vendor data produces the answer “reduce Vendor B's allocation by 40% and reallocate to Vendor C,” that is intelligence. If it produces “our marketing performance was mixed,” that is data dressed up as a conclusion.
Building Intelligence, Not More Data
The shift from data to intelligence requires three things that most PI firms have not yet built.
First: connected data. Marketing spend, intake outcomes, and case results need to live in a single view — or at least be systematically connected — rather than in separate systems that must be manually reconciled. Intelligence emerges from connections. Disconnected data sets produce disconnected insights at best and no insights at all at worst. The RevenueScale platform exists specifically to create this connected view for PI firms.
Second: the right questions as design constraints. Building a measurement system starts with the decisions you need to make — vendor allocation, intake improvement, budget justification — and works backward to identify which data points must be connected and how they should be displayed to answer those questions. Too many firms build measurement systems around the data they happen to have, then wonder why the reports do not produce clear decisions.
Third: continuous monitoring rather than periodic reporting. Intelligence that is two months old when you receive it is often too old to act on. The firms that move from monthly reports to dashboards with continuous updates describe a fundamental shift in how they manage marketing — from reactive to proactive, from retrospective to current. AI-powered monitoring is how leading PI firms make this shift without adding analyst headcount.
Data collection is not the problem for most PI firms. The problem is converting data into intelligence. The firms that have solved this problem are not managing more metrics — they are managing fewer, better ones. And those metrics are connected in ways that produce clear decisions rather than more information to absorb.