How to evaluate a robot activation without vanity metrics
Write the decision the event is meant to inform before choosing metrics. A clear objective helps the team separate what it did from what guests learned, experienced, or chose to do afterward. Select a collection method that fits the question, record the conditions around the data, and report the result with its limits. A count can describe an activation; it cannot, by itself, prove sales, brand lift, or return on investment.
Write the objective before choosing metrics
Begin with one sentence: “We are running this activation so that [audience] can [observable outcome] during [event context].” For example, a technology conference might want attendees to identify one task demonstrated by the robot. A community event might want visitors to find the activity and receive an accessible explanation. These are planning examples, not established predictors of business results.
Ask what decision the finding will support. Will the team use it to improve the next run, decide whether a format fits the audience, or explain event activity to a sponsor? If no one can name a decision, reconsider collecting the information. Define the audience and time window before selecting a measure; otherwise, a convenient number can end up standing in for a question it does not answer.
The European Commission’s communication-evaluation toolkit frames evaluation around assessing activities and supporting improvement. A scoping review of 44 conference-evaluation studies describes varied objectives and practices, which is a reminder to define the purpose and method for the event at hand rather than assume one metric works everywhere. European Commission toolkit · Conference evaluation scoping review
Separate activity from outcomes
Activity measures describe what the team delivered: demonstration starts, session attendance, questions received, or time the activation was available. Reach measures describe who encountered it, if that can be observed responsibly. Outcomes concern a change relevant to the objective, such as whether guests can identify the demonstrated task immediately afterward. Longer-term outcomes require a suitable follow-up and a way to interpret what else may have influenced the result.
Keep those categories distinct in the report. “We ran six demonstrations and spoke with 42 visitors” describes delivery and interaction. It does not establish that 42 people understood the technology, remembered the event, changed their opinion, or purchased anything. “Most of the people who answered our optional exit question could name the task” is more specific, but still describes respondents rather than every attendee. Do not convert engagement into brand lift or revenue without a method designed to support that inference.
Choose a collection method
Match the method to the objective and the pace of the room. A staff tally can record counts with little interruption. A brief voluntary question can check immediate understanding. A short survey can gather reactions, while a staff debrief can capture operational issues that a guest metric will miss. Each method answers a different question. Decide who collects the information, when, where it is stored, and how the team will avoid recording information it does not need.
Keep prompts neutral and easy to answer. “What is one task you saw the robot demonstrate?” is more useful for recall than “Did you enjoy our amazing robot?” If you ask about experience, offer a range of answers and a way to skip. Do not collect names or contact details unless they are necessary for a separately explained purpose and the event has an approved process for handling them.
Pilot the collection method during rehearsal. Check whether staff can use it while handling their assigned role and whether the wording works in the actual sound and traffic conditions. If collecting data would distract the operator or block the guest route, change the timing or use a lighter method.
Record limits and context
Log the date, venue area, time window, demonstration format, number of sessions, relevant schedule changes, and collection method. Record who was invited to respond and whether responses were voluntary. Note interruptions, unavailable sessions, room changes, or staff shortages that may affect interpretation. A small set of contextual details often explains a result better than a polished chart without a field note.
State the denominator whenever you can. If 18 of 25 respondents selected an answer, report both numbers and explain how they were recruited. Avoid presenting the proportion as representative of all event guests unless the collection design supports that claim. Do not compare two events as if they were equivalent when audience, schedule, venue, and method differed materially.
Share findings responsibly
Use a compact report: objective, method, activity, outcome observation, limitations, and next decision. Separate direct observations from interpretations. “The queue reached the marked holding point twice” is an observation. “The format was a success” is a conclusion that needs a defined success criterion and supporting evidence.
Share what the data can support and preserve the caveat beside the headline result. If the measure only describes activity, call it activity. If the sample is self-selected, say so. If the team lacks a baseline or comparison, do not claim the activation caused a change. Use the findings to improve the next brief, layout, explanation, or collection method—not to manufacture certainty from a memorable moment.
Method and limits
This worksheet applies general evaluation principles to a one-off event activation. The cited toolkit and conference review provide methods and examples, not validated predictors for robot events. Candidate measures here do not prove sales, brand lift, or ROI. These prompts do not certify legal, accessibility, or safety compliance and do not predict audience or business outcomes.
Sources
- European Commission, “Toolkit for the evaluation of the communication activities”: https://commission.europa.eu/system/files/2019-10/communication-evaluation-toolkit_en.pdf
- Justin Neves, “A scoping review about conference objectives and evaluative practices: how do we get more out of them?”: https://pmc.ncbi.nlm.nih.gov/articles/PMC3487916/