Pillar · Reputation
The average is what gets reported. The spread is what gets experienced.
Mapboard does not roll your network up into one number. It collects what people say about each Location, classifies it by topic, and shows you which one is sliding before it shows up in the average.
“Got there at 9 and it was closed, even though Maps says it opens at 8. Second time this has happened to me at this Location.”
Orbit suggests a reply
We're sorry this happened. The published hours have been corrected and we're reviewing the opening routine at that Location. Thank you for flagging it both times.
01 · The problem
A network averaging 4.0 has Locations sitting at 2.5
The average is the number that goes into the report and the one that makes the board feel good. But nobody visits the average: the customer walks into one specific Location, and if they got the 2.5, that is your brand to them.
Reputation tools make the problem worse. They roll reviews, surveys and visibility into a single score, and that score hides exactly what you need to find: the Location that is sliding while the aggregate does not move.
02 · How it works
From what they say to what you do
Five steps. The last one is what turns an opinion into work with an owner.
- 01
It gets collected
Google reviews arrive in real time, not in an overnight sweep. Along with public comments and recommendations from your pages.
Every review stays attached to its Location. When a social signal does not carry a Location, it is attributed by exact match or left flagged as a candidate: we never invent which one it belongs to.
- 02
It gets understood
AI classifies every review by sentiment, topic and severity, so a Location's pattern becomes visible instead of one loose complaint.
On top of that you can tag by reason in your own vocabulary, and see reason cross-referenced against Location. That is where it becomes clear that it is not "bad service", it is the checkout at one specific store on Saturdays.
- 03
It gets answered
AI drafts in your brand's voice and your team edits and approves before it publishes. The text that goes out is the text a person approved.
For the repetitive part there are template auto-replies, with rules by star range, a waiting window and a cutoff from the moment you switch them on: it never answers your back catalog all at once. And you can always preview what it would reply before enabling it.
- 04
It gets compared
Each Location against its real peers, same format and same area, and against the competitors that opened on its own block.
Rating gap against the immediate surroundings, review share, local position and which brands show up near how many of your Locations. Comparing your network against a national average says nothing; against the store across the street, it does.
- 05
It becomes work
When a review stops being an opinion and is a problem, it turns into a case with an owner, a severity and a clock.
AI proposes and a person approves. That is where reputation ends and operations begin, which is where the problem actually gets fixed.
You will not find a reputation score.
It was removed on purpose. A number that blends reviews, visibility and completeness produces false precision and hides the variance. In its place there is a state per Location, with the rules in plain view: you can ask why a Location is in that state and see what was evaluated.
- Incorrect hours62
- Checkout wait time48
- Out of stock31
- Staff attitude19
The most repeated reason concentrates in 9 of 1,492 Locations.
03 · The difference
What a score will never tell you.
The aggregate is the enemy, not the goal
Every tool in the category rolls your network into one KPI. If what you are looking for is the Location that is sliding, that number works against you: it moves so little it never notices. Mapboard disaggregates by design.
Fair comparisons, not vanity rankings
A highway store does not compare to a mall store. Nested groups let you measure each Location against its real peers, and competitive intelligence measures it against whoever actually takes its customers: the business across the street.
Data honesty as a product trait
When an analysis was run on a sample, the platform says so on screen. When a Location has few reviews, the window widens on its own rather than classifying it off three opinions. We would rather give you a number with an asterisk than a pretty one that is false.
Downtown Brooklyn
- This Location3.2
- Its peers, same format4.3
- Competitors within 1 mile4.1
A full point below its peers. In the network average, invisible.
04 · Made to fit
It adapts to how your customers talk
A pharmacy's vocabulary is not a fuel station's, and your brand's tone is nobody else's.
Reason tags
Your own taxonomy of what people complain about, with cross analytics of reason by Location.
Voice guide
The tone the AI replies in, defined once and applied across the network.
Reply templates
For the repetitive part, with rules by star range and a waiting window.
Scope by group
Which rules apply to which areas or formats, without it being all or nothing.
Email reports
Reviews by reason, with a spreadsheet attached and the detail of what fell inside and outside the cutoff.
QR per Location
To ask for reviews at the counter, with a scan count per store.
From Google in real time, and from Meta the reviews and recommendations on your pages. Replying happens on Google: Meta does not allow answering reviews through its API, and we would rather tell you than have you find out in the demo. Apple Maps and Bing are not there today.
It drafts, it does not publish. Your team edits and approves, and the text that goes out is the one a person approved. Separately there are template auto-replies for the repetitive part: those are texts you wrote and approved beforehand, with rules for when they apply, and you can preview what it would reply before turning it on.
Because it blended dimensions that are not comparable to each other and implied a precision that did not exist. A 78 out of 100 tells nobody what to do on Monday. In its place there is a state per Location with the rules in view, and the spread across the network, which is what actually points at where to act.
Against their peers inside your network, grouped by format or area, and against the real competitors in their immediate surroundings: who opened nearby, what rating they hold, how many reviews they gather and where you land on that block. It is a capability enabled per customer.
They are not classified off three opinions. If a Location has low volume in the short window, the analysis widens to a longer period so the result means something. It is built for low-volume networks, like neighborhood pharmacies.
It becomes a case: owner, severity, SLA clock and activity log. The AI proposes it and a person approves. Answering a review about a dirty store well does not clean the store; opening the case does.
Let's find where your network is sliding.
You tell us how you measure the reputation of your Locations today and we show you, on a demonstration network, what the spread looks like: which Location is dragging the average down, and for what reason.
