Stage 1: a relevant signal
Identify the specific event that shows someone may want a conversation.
In this article
The situation
A visit, impression or generic reaction is not automatically a lead. Define which observable event starts the response clock for each channel and keep its source attached to the record.
Define the entry event precisely
A signal is an observable action that may deserve a response, such as a direct question, a relevant comment, a form submission or a call. Do not collapse it into impressions, followers or visits. List the channels that actually produced inquiries last month and classify which events were actionable. If the source is unknown, map it before evaluating a new channel.
Use a consecutive sample rather than selecting only opportunities that became customers. The first stage should tell the team what arrived, when and through which route, while preserving an unknown category for missing records.
What changes the next decision
The source affects the expected answer and contact basis. A public social comment may call for a short public answer, while a form requesting a quote may need a private, researched reply. Carry the original context forward so the responder does not send the same generic message to every source.
A stage-one repair is often simple: name the primary source, record it consistently and separate relevant signals from general engagement. Only after the team can do that should it decide whether a product could improve the response.
A practical stage-one log
Record the event type, source, arrival time, whether a person expressed a question and the first owner who could act. Separate a direct request from passive engagement. A visible unknown option is preferable to guessing a source because it shows where tracking or staff notes need improvement.
Review the distribution over a defined period, then choose one primary path to map end to end. Do not combine unrelated channels merely to produce a larger total. The next stage can only collect the right context when the team understands how the person arrived.
Work through it
- Name the entry event separately for social, website, phone, content and outbound. Write one example that qualifies and one that does not.
- Count relevant signals for a recent, defined period. Keep the original channel and date so later results are not detached from their source.
- Choose one signal to improve first; a small, clear sample is more useful than a combined total with no context.
What to measure
- Relevant signals by source
- Signals with a known arrival time
- Signals that reach a named responder
Boundaries
- Do not compare a social reaction with a booked call as equivalent units.
- Avoid industry benchmark claims until the site has a real sample and method.
Questions for your own workflow
- What exact event starts the clock?
- Which source produced it?
- What would make it irrelevant?