Type what you're actually trying to do - "a coffee shop somewhere busy in Houston" - and Scout Bee sets up the map, shows you exactly what it changed, and lets you undo it. Here is everything it can do, using Houston throughout.
Most location tools ask you to know the vocabulary before they will help you: pick a dataset, pick a variable, pick a classification method. Scout Bee works the other way round. You type what you are actually trying to do - “I want to open a coffee shop somewhere busy in Houston” - and it builds the map for you, then shows you exactly what it changed so you can argue with it.
This guide walks through everything Scout Bee can do, using Houston for every example. There is no jargon you need in advance; each term is explained the first time it shows up, and there is a glossary at the end. If you have never opened a mapping tool in your life, you are the reader this was written for.
One sentence in the chat panel on the right; a scored map of the whole of Houston on the left. Nothing else was clicked.
First, what the map actually is
MapBees does not think in streets or postcodes. It chops a city into thousands of small, equal-sized six-sided tiles - hexes - each roughly the size of a couple of city blocks. Houston is 49,533 of them. For every single tile it already knows things like how many people live there, how many coffee shops or offices or schools sit inside it, how much building space there is, and how well served it is by transit.
That is the whole trick. Because every tile carries the same set of numbers, any question of the form “where is there a lot of this and not much of that?” becomes something a computer can answer for the entire city at once. Scout Bee's job is to turn your sentence into that question.
In plain English: Hex
A small tile on the map. Everything - people, shops, buildings - is counted tile by tile, which is what lets you add up any shape you like later.
Houston before you ask anything: 49,533 tiles, each already carrying its own numbers.
How a conversation works
Scout Bee lives behind the bee button at the bottom-right of the map. Open it and you get a chat box with a few suggestions. You are not picking from a menu - anything you can describe is fair game.
The starting point. Type a sentence, or tap one of the suggestions.
Every time Scout Bee changes something, it shows a Map setup card listing each step in plain English. Three things sit under that card, and they are the reason you can trust it:
On the map - the change has already been applied. (You can switch Auto-apply off in the header if you would rather approve each one yourself.)
Undo - puts the map back exactly as it was. Nothing Scout Bee does is one-way.
Copy link - a link that reopens this exact map for anyone you send it to.
The important bit
Scout Bee never hides what it did behind a black box. Whatever it sets up appears in the same panels and sliders you would have used yourself - and you can drag any of them afterwards.
1
Scoring: “where should I open?”
This is the classic question, and it is what the score is for. A score is a 0-100 rating given to every tile in the city, worked out from the things you said matter. Here is the request, typed exactly as a person would say it out loud:
I want to open a specialty coffee shop in Houston. Where should I look? I want plenty of people around, lots of offices nearby, and not too many coffee shops already.
Scout Bee reads that as three separate instructions, finds the matching data for Houston, and reports back what it is about to do before doing it:
The plan, then the Map setup card: three signals, their weights, and which direction each one points.
Read the card slowly, because it contains the whole idea. It says the map is scored by Total (residents) at 40%, Corporate Office at 40%, and Coffee Shop at 20% - and the first two are marked high while the last is marked low.
In plain English: Cluster and Gap
Every signal points one of two ways. Cluster (“high”) means find me lots of this - you want customers, so more residents is better. Gap (“low”) means find me where there is little of this - you do not want to open next to four existing coffee shops. Same data, opposite direction.
Those percentages are the weights, and they always add up to 100%. They are the honest answer to “how did it decide?”: a tile with lots of residents and offices but three coffee shops already will score lower than an equally busy tile with none.
What Scout Bee wrote into the ordinary settings panel. Every slider is yours to drag; the map re-scores as you move it.
On the map itself, brighter yellow means a better match and dark purple means a worse one. If you want to know why a particular tile scored what it did, hover it: you get the score out of 100 and the raw numbers behind it, with no interpretation in between.
Hover any tile for its score and the raw counts that produced it.
2
Changing your mind costs one sentence
The first answer is rarely the last one. Rather than hunting for the right slider, you say what changed - and Scout Bee keeps the rest of the conversation in its head:
Make the offices signal count twice as much as the other two, and rank the top 10 areas.
It does not ask which offices, or which two, because it already knows. Offices go to 50% with residents and coffee shops at 25% each, and the ten best tiles get numbered badges on the map so you can go and look at them one by one.
A follow-up, not a new question. The criteria carry over; only the weighting changes.The same city, re-scored. Numbered badges mark the ten strongest tiles under the new priorities.
In plain English
Weights are relative, so you can talk about them the way you would to a colleague: “twice as important”, “mostly families, a bit of footfall”, “forget the transit part”. You never have to work out the percentages yourself.
3
Bivariate: two things at once
A score blends everything into one number, which is exactly what you want for a shortlist and exactly what you do not want when you are trying to spot a mismatch. That is what a bivariate map is for: it shades every tile by two measures at the same time and leaves them separate, so the interesting corners stay visible.
Show me a bivariate map of Houston: where people live on one axis, and how much retail building space there is on the other, so I can spot places with lots of residents but little shopping space.
Two measures, one map - and Scout Bee says which one went on which axis.
The key is a three-by-three grid of colours. Move right for more of the first measure, up for more of the second. So the pale bottom-left corner is “not much of either”, the dark top-right is “lots of both”, and the two off-corners are the ones worth staring at: plenty of residents, very little retail space - or the reverse.
The key. Right for more population, up for more retail space - and the corners are where the stories are.Houston by residents against retail space. The teal spread is where people live without much shopping nearby; the purple veins trace the commercial corridors.
In plain English: “But the units are different!”
They are - people versus cubic metres of building - and it does not matter. Each axis is ranked only against itself, so the two are never added together. You are reading the pattern, not the arithmetic.
4
Territories: who is actually within reach?
City-wide maps are for finding candidates. Once you have a specific address, the question changes to “who can actually get here?” A territory answers that: a zone drawn around a point along real streets - everywhere within a 15-minute walk, a 10-minute drive, two kilometres, whatever you say - with everything inside it totalled up.
I'm scouting two spots in Houston: 6100 Main Street and 1500 McKinney Street. Draw a 15-minute walking area around each and tell me how many people live inside.
Scout Bee looks up both addresses, draws both walking areas, adds up the residents inside each, and answers the question in the chat before you have opened a single panel:
Two addresses in, two answers out: about 5,490 residents within a 15-minute walk of one, 9,276 of the other.
That gap is the entire point. Two Houston addresses a little over three miles apart, both perfectly respectable - and one has nearly 70% more walk-up population than the other. Notice too that the shapes are not circles: they follow the street network, so a freeway or a bayou that nobody walks across genuinely cuts the area short.
The two 15-minute walking areas, drawn along real streets rather than as circles.
Every area you create is listed in the Areas panel with its travel time and size, so you can rename, hide, or delete any of them - and add more to compare.
The Areas panel. Scout Bee filled it in, but every control is the ordinary one you would have used.
Open any area's full dashboard and you get everything inside it, counted: the area in square kilometres, the number of named places, the total building space, and a breakdown of what kind of buildings they are. For the 6100 Main Street walking area that is 2.9 km², 1,299 named places, and 10.6 million cubic metres of building - led by medical and education space, with retail near the bottom. That tells you what sort of neighbourhood this is before you have visited it.
The dashboard for one walking area: size, places, building space, and what kind of buildings they are.
Scroll on and the same area is broken down by people rather than buildings: age bands, household types, marital status, families with children, and the local population mix - all from public census data, all for the exact shape you drew.
The same walking area, described by its residents instead of its buildings.
A note on the numbers
These come from public census and open map data, aggregated to tiles. They are excellent for comparing one place against another, and they are not a substitute for standing on the corner at 8am on a Tuesday.
5
Market share: how much of it would I win?
Territories tell you who is nearby. Market share tells you how those customers split when your competitors are reaching for the same people. You give Scout Bee your locations and theirs, say how far customers will travel, and it works out who wins whom.
I run two coffee shops in Houston, at 1200 McKinney Street and at 2800 Kirby Drive. My two competitors are at 500 Dallas Street and at 3400 Montrose Boulevard. Put all four on the map, give every one a 10-minute driving trade area, and run the market share analysis.
That single sentence becomes a four-step plan: switch to the market tool, place four pins, give them all a 10-minute driving trade area, and run the numbers.
Four steps, each ticked off, ending with the headline: your stores capture about 55% of the demand in play.
On the map, each shop gets its 10-minute driving trade area. Where two of those overlap, the tiles are shaded by who is likely to win the customers living there - blue shades for your shops, red and orange for the competition.
Four 10-minute driving trade areas over central Houston, shaded by who is likely to win the overlap.
Then the report puts a number on it:
55.0%
My market share
122,808
Customers my shops capture
223,236
Total demand in the trade areas
And it breaks that down shop by shop. “Exclusive” customers are the ones only that shop can reach; “contested” are the ones inside more than one trade area, split between the shops competing for them.
Location
Whose
Exclusive
Contested
Captured
Share
2800 Kirby Drive
Mine
41,168
35,830
76,997
34.5%
3400 Montrose Boulevard
Competitor
7,331
43,721
51,052
22.9%
500 Dallas Street
Competitor
14,703
34,672
49,376
22.1%
1200 McKinney Street
Mine
11,144
34,666
45,811
20.5%
Read across the top row and the story writes itself: the Kirby Drive shop wins 34.5% of the market almost entirely on exclusive customers - 41,168 people no competitor in this analysis can reach. The McKinney Street shop, downtown and hemmed in, wins 20.5% and has to fight for most of it. If you could only keep one, you now know which.
The full report: your share, the split, and a ranked line for every location in the analysis.
Two more numbers are worth knowing. Contested demand - 148,889 people here - is how much of the market is genuinely up for grabs rather than locked in. And cannibalization - only 863 people in this case - is how many customers your own two shops are taking from each other. A low number means they are well spread; a high one means you have opened next to yourself.
The report also describes the contested zone as a place in its own right: 46.2 km², 25,226 named places, and a full demographic profile of the people living in it. That is the ground you and your competitors are actually fighting over.
Who lives in the contested area - the ground genuinely in play, profiled like any other territory.
Decision support, not a forecast
Market share here is a model of pull - closer and stronger locations win more - not a sales prediction. Use it to compare options and spot the shape of a network, and expect reality to have opinions of its own.
6
What it deliberately will not do
Scout Bee will not invent your locations. It geocoded four Houston addresses above because you gave it four Houston addresses. Ask for a market-share analysis without naming any, and it does something more useful than guessing:
I'm thinking about a bakery in Houston and I want to check market share, but I'd rather drop the pins myself.
No addresses given, so it sets everything up it can reason about - then stops and hands you the map.
It picks a sensible trade area for the business type - a bakery gets a short walking catchment, not a ten-minute drive - opens a fresh project so your existing work is untouched, arms the map for clicking, and waits. You drop your own pins, tell it you are done, and it runs the analysis on real locations.
Your pins, placed by you. Scout Bee did the setup; it did not pretend to know where your bakery is.
Why this matters
Where your shops are, and who you consider a competitor, is knowledge only you have. An assistant that guesses produces a confident, professional-looking report about a business that does not exist. This one asks.
7
Every answer is a link
Notice the URL in each of those chat screenshots. Every setup Scout Bee builds is written into a link that rebuilds the whole thing - signals, weights, drawn areas, store pins, the lot. Paste it to a colleague and they open your map, not a screenshot of it. They can then change one weight and send it back.
That also means you can keep working by hand at any point. Scout Bee is a faster way into the map, not a separate mode you get locked into - the tool panels stay exactly where they were, and so does everything you had already set up.
Sentences you can copy
Swap in your own city and business. The first five produced the figures above, word for word; the last is simply a good way to find out what a city has before you ask for anything.
Site selection
“I want to open a specialty coffee shop in Houston. Where should I look? I want plenty of people around, lots of offices nearby, and not too many coffee shops already.”
Adjust what matters
“Make the offices signal count twice as much as the other two, and rank the top 10 areas.”
Two things at once
“Show me a bivariate map of Houston: where people live on one axis, and how much retail building space there is on the other.”
Walkable areas
“I'm scouting two spots in Houston: 6100 Main Street and 1500 McKinney Street. Draw a 15-minute walking area around each and tell me how many people live inside.”
Market share
“I run two coffee shops in Houston, at 1200 McKinney Street and at 2800 Kirby Drive. My two competitors are at 500 Dallas Street and at 3400 Montrose Boulevard. Put all four on the map, give every one a 10-minute driving trade area, and run the market share analysis.”
Just asking
“What can MapBees tell me about Houston?”
Where it still gets things wrong
Scout Bee is marked Beta in the panel, and that is not decoration. Three honest limits:
It sometimes picks a near-miss signal. Ask for “offices” and you might get a specific office category rather than the one you pictured. The Map setup card names exactly what it chose, so check it - and tell it to swap.
Address lookup prefers full addresses. “6100 Main Street, Houston” is reliable; a landmark name on its own is less so. Give it a street address when precision matters.
Ask again if a reply looks thin. The same question can occasionally come back as prose with no map attached. Rephrasing, or simply asking again, usually fixes it - and Undo means a wrong turn costs nothing.
None of these are dangerous, because nothing is hidden: the card says what changed, the panels show it, and Undo puts it back.
Every term, in one place
Bookmark this if a word ever trips you up.
Scout Bee
The assistant inside the MapBees map. You describe what you are looking for in ordinary words, and it sets the map up for you.
Hex (tile)
The map is chopped into small six-sided tiles, each about the size of a couple of city blocks. Every number - people, shops, buildings - is counted tile by tile.
Signal
One thing that can be counted in a tile: residents, coffee shops, offices, transit stops, building space, and so on.
Score
A 0-100 rating for each tile, worked out from the signals you said you care about. Brighter tile, better match.
Weight
How much one signal counts compared with the others. Doubling a weight means that signal matters twice as much.
Cluster vs Gap
Cluster means 'find me lots of this'. Gap means 'find me where there is little or none of this'. Every signal is set to one or the other.
Bivariate map
One map shaded by two things at once, so you can see where they line up - and, more usefully, where they do not.
Territory (area)
A zone drawn around a place - for example everywhere within a 15-minute walk - so you can total up who and what is inside it.
Catchment / trade area
The area a shop can realistically serve. MapBees draws it along real streets, not as a circle.
Market share
Of all the customers in the area you analysed, the percentage your locations are likely to win.
Contested demand
Customers who sit inside more than one trade area, so several shops could plausibly win them. They get split rather than handed to one shop.
Cannibalization
Customers your own locations take from each other because they sit too close together.
The whole thing, in six lines
Open the bee at the bottom-right of the map and describe what you are looking for.
Scout Bee finds the right data for your city and shows a Map setup card before anything changes.
Scoring ranks the whole city; bivariate compares two measures; territories total up one area.
Market share splits real customers between your locations and your competitors'.
It never invents your store locations - if you have not given them, it hands you the map.
Undo reverses anything, and every setup is a link you can share.
You do not need to know what a hex, an isochrone or a gravity model is to use any of this. You need a sentence about what you are trying to do.
Scout Bee is the assistant built into the MapBees map. You describe what you're looking for in ordinary words and it sets up the map for you - choosing the data, the weighting, and the tool - then shows exactly what it changed.
Do I need any mapping or GIS experience to use it?
No. You type a sentence such as "I want to open a coffee shop in Houston where there are plenty of people and offices but few coffee shops", and Scout Bee builds the map. Every term it uses is visible in the panel and explained on the page.
Can Scout Bee do market share and territory analysis, not just scoring?
Yes. It drives four workflows: scoring a whole city for site selection, comparing two measures on one bivariate map, drawing travel-time territories with full demographics, and running a market-share analysis against your competitors.
Will Scout Bee guess where my stores and competitors are?
No, and that is deliberate. It places pins only for addresses you give it. If you ask for a market-share analysis without naming locations, it prepares the workspace, arms the map for clicking, and hands the pin placement to you.
Can I undo what Scout Bee changes, or share the result?
Both. Every change comes with an Undo that restores the previous map exactly, and a Copy link that rebuilds the same map - signals, weights, areas and pins - for anyone you send it to.
Are the numbers a forecast?
No. They are decision support built on public census and open map data - excellent for comparing and shortlisting places, and meant to be confirmed with local knowledge before you sign a lease.
See it on a real map
The best way to understand the score is to play with it. The full app is on a free tier, with no sales call.
Builds MapBees's location-intelligence pipeline end to end — census and Overture Maps ingestion, H3 hex scoring, and the map interface — and writes the methodology behind every number on the site.
