A buyer may see a news story, return through search a week later, ask a colleague, and then respond to a sales email. No single click tells the complete story. PR attribution is therefore a question of evidence and assumptions: which interactions were observed, what changed after the activity, and which alternative explanations remain? A credible answer can show PR's contribution without claiming every later sale was caused by a placement.
Define the outcome and the observation window
Choose a business result that the organization actually records: qualified inquiries, trial starts, meetings, completed orders, or revenue from a defined customer cohort. State the time period and geography. A national story may have a long consideration cycle, while a local event announcement can have a short response window. Match the window to the buying process rather than choosing it after seeing the result.
Record a baseline for the same measure before the activity. Note other campaigns, price changes, product launches, holidays, and changes in sales capacity. These factors can influence the same outcome and should appear in the report, not be treated as noise to ignore.
Capture direct signals first
01
Referral visits
When a publication links to the business, web analytics may show visitors arriving from that page. Check the referring source, landing page, and subsequent actions; a visit is not a sale.
02
Tagged links you control
Use consistent UTM parameters for links in owned emails, partner messages, or paid amplification. Google Analytics documents how campaign parameters populate acquisition reports. Do not imply that an independent publisher will preserve tags you did not control.
03
Sales conversations
Add a neutral 'How did you hear about us?' question and capture volunteered answers in the CRM. Let buyers mention more than one influence. Train staff to record the answer rather than infer it from the latest campaign.
04
Branded demand
Compare branded search or direct inquiries before and after major coverage, but label the result as an association unless the analysis controls for other causes.
Move from association toward stronger evidence
A holdout geography, matched audience, or carefully timed comparison can provide stronger evidence than a simple before-and-after chart. These designs take planning and may not fit small campaigns. Surveys can ask whether people recall a specific story, but recall and stated influence are not the same as actual purchasing behavior.
Multi-touch models assign credit among recorded interactions; their output depends on the model's rules and the touchpoints it can see. Untracked offline conversations and cross-device behavior remain outside the data. Show modeled PR-influenced revenue separately from orders with a directly observed referral, and never sum overlapping categories as if each were an independent sale.
Report the result with limitations
A concise report should show the PR activity, dates, relevant coverage, measured visits or inquiries, qualified leads, closed deals if available, and total program cost. Identify the source for each number and state whether it is directly observed, self-reported, or modeled. A conversion rate needs a numerator and denominator from the same cohort and period.
AMEC's evaluation guidance helps separate publication outputs from audience outcomes and business impact. Use the evidence to improve story selection and follow-up, not to manufacture a precise return-on-investment figure from impression counts. If revenue data or deal cost is unavailable, say so plainly and report the strongest observable result instead.