Dwell time is the time a recorded visit spends in a defined area. For a store or shopping mall, that area might be a department, a queue, a corridor or the entire premises. A useful report states the area, the observation period and how a visit begins and ends. A number without those definitions is difficult to compare or act on.
Longer dwell is not automatically better. A visitor may be enjoying a display, waiting for help or standing in a slow queue. The decision should depend on the purpose of the space and the other evidence available.
Start with one operating question
Choose a question that a team can act on. Examples include: Are shoppers finding the demonstration area? Does the fitting-room queue need a different staffing pattern? Does an event send visitors into nearby stores, or keep them in the atrium?
Then identify the decision owner. A store manager may be able to change queue staffing immediately; a tenant-mix question needs a different process. Avoid collecting a large set of behavioural measurements before anyone knows which decision they will support.
Write a measurement definition
A practical specification can include:
- Zone: the exact boundary and the version of the floor plan.
- Visit start: the event that begins the observation.
- Visit end: the event that ends it, including any inactivity rule.
- Repeat entry: whether leaving and returning creates a new visit.
- Eligible observations: which incomplete, staff or test records are excluded, if the system can reliably identify them.
- Time window: opening hours, days and time zone.
- Coverage: which entrances or areas are actually measured.
These are proposed reporting fields, not a claim that every sensor or Fundle installation already supports them. If a source cannot distinguish repeat visits or employees, state that limitation rather than silently adjusting the numbers.
Use a distribution, not only one average
Consider five hypothetical recorded visits lasting 2, 3, 4, 6 and 25 minutes. Their average is 8 minutes: 40 minutes divided by five visits. Their median is 4 minutes: the middle value after sorting.
Both are correct, but the average is pulled upward by one long visit. Showing the visit count, median and average makes the pattern easier to interpret. For a larger dataset, agreed duration bands or percentiles can add context. An outlier should be investigated, not automatically discarded simply because it changes the result.
Also avoid averaging averages across unequal groups. If one zone has ten visits and another has a thousand, a simple average of their two averages does not represent the average visit across both zones. Use total recorded duration divided by total eligible visits when that combined definition is meaningful.
Read time alongside the purpose of the space
For a product demonstration area, longer visits might justify checking whether people are interacting with the display. For a payment queue, shorter waiting time may be the desired outcome. A busy corridor may need clear passage rather than an activation that encourages people to stop.
Pair the dwell report with relevant evidence: observed queue length, staff notes, product availability, transactions or the event timetable. Keep the units explicit. A zone visit is not necessarily a unique customer, a transaction or a loyalty member.
Check the data before changing the store
Review a small set of observed periods with the team responsible for the measurement. Look for missing intervals, clock differences, shifted zone boundaries and coverage changes. Check whether a layout change has affected the measurement itself.
For example, moving a display into a sensor's field of view could change what is observed as well as how people move. If a report suddenly improves on the same day a camera or rule changes, investigate the measurement change before attributing the improvement to customer behaviour.
Compare like with like
A weekday morning and a festival weekend have different conditions. Record promotions, closures, events and unusual operating hours. Where a credible experiment is feasible, plan the comparison before changing the layout, not after selecting the best-looking result.
The NIST experimental-design handbook emphasises defining objectives and choosing variables and a design. Applying that discipline to retail means deciding what will change, what outcome matters and what alternative explanations need to be considered. The exact retail test still needs its own design. NIST: Choosing an experimental design
A small pilot brief
Here is an illustrative brief, not a customer case study:
- Question: Is the service desk difficult to find?
- Change: Test clearer wayfinding without changing the service offer.
- Primary outcome: An agreed measure of successful arrival at the desk, if measurable.
- Supporting observations: Visits to nearby zones, queue waiting, staff questions and complaints.
- Comparison: Similar operating periods or locations, with major differences documented.
- Stop condition: Reconsider the change if it obstructs movement or creates another operational problem.
The pilot does not need a promised percentage uplift. It needs a decision that the evidence can support.
Connect dwell and sales carefully
Aggregate zone activity and sales may move together without one causing the other. Do not label every shopper present near a store as a buyer, or join movement records to named loyalty members merely because both datasets exist. Define the business purpose, permissions and access arrangements before combining data.
Fundle's mall sales workflow is a separate place to discuss tenant-sales reporting. Treat the linkage between movement and sales as a data-design question, not an assumed feature of a heat map.
Common questions
What is a good dwell time?
There is no useful universal target across a queue, a supermarket aisle and a leisure zone. Start with the purpose of the area and a consistent local baseline.
Can dwell time prove that an event increased revenue?
No. It is one observation. Revenue attribution requires an appropriate comparison and a clear connection between the measurements.
To scope a practical measurement discussion, request a Fundle working session. Bring one floor plan, one decision and an inventory of the data you already have.