Location-based spend analysis

Location-based spend analysis is the practice of grouping total expenditure by geographic dimension, such as country, region, city or individual site, to see where money is being spent and why. If a business spends 4 million pounds across twelve sites, the analysis breaks that figure into a per-site number, so you can see that one site accounts for 900,000 pounds while another spends 110,000 pounds. The headline total stays the same, but the distribution underneath it becomes visible.

The value of the analysis is comparison. Once spend is mapped to place, you can rank locations, compare them against a per-location budget, and normalise for size so a large site and a small site can be judged fairly. A site that spends more in absolute terms is not automatically a problem. The signal is spend that is high relative to the headcount, revenue or square footage at that location.

Location-based spend analysis is most useful for organisations with physical footprint or regional structure: retail estates, manufacturing plants, multi-office service firms, and field operations. It is the geographic complement to category and vendor analysis. Category analysis tells you what you are buying, vendor analysis tells you who you are buying from, and location analysis tells you where the money lands.

The core calculation attributes every line of expenditure to a location, sums spend per location, and expresses each location as a share of the total. To make locations comparable, the raw figure is then normalised against a denominator that reflects the size of the site, such as revenue, headcount or floor area.

1. 1

   Tag every transaction with a location

   Assign each invoice, card charge and expense claim to the site, region or country that consumed it. Transactions with no location tag are the single biggest source of unreliable analysis, so unassigned spend should be tracked as its own bucket and driven down over time.

2. 2

   Sum spend per location

   Aggregate the tagged transactions to a clean total for each location in the period. Keep the underlying detail so a high-spend site can be drilled into by category and vendor without re-running the whole calculation.

3. 3

   Calculate each location share

   Divide spend at each location by total spend across all locations to get its percentage of the whole. This shows concentration: whether spend is spread evenly or dominated by a handful of sites.

4. 4

   Normalise for size

   Divide location spend by a size measure such as revenue, headcount or square footage. A site spending 900,000 pounds on 200 people (4,500 pounds per head) may be more efficient than one spending 300,000 pounds on 40 people (7,500 pounds per head).

A worked example makes the difference clear. Two sites both spend 600,000 pounds on facilities. Site A houses 300 staff, Site B houses 90 staff. On the headline number they look identical. Normalised per head, Site A spends 2,000 pounds and Site B spends roughly 6,700 pounds. The analysis has flipped a tie into a clear outlier, and Site B is now the place to investigate.

A metric tree decomposes total spend by geography first, then by the cost drivers within each location. This turns a flat ranking of sites into a diagnostic structure, because every branch points to a specific lever and a specific owner.

The first level splits total spend into regions, and each region splits into the sites beneath it. Each site then decomposes into the cost categories that make it up, such as facilities, payroll-adjacent spend, logistics and local procurement. The leaf level is where the drivers live: lease rates, energy use, supplier pricing and local maverick spend that bypasses agreed contracts.

The tree lets you trace a movement back to its cause. If regional spend rises, the tree shows whether the cause is one site, one category within a site, or a price change that hit every location at once. A central energy price rise looks different in the tree from a single site quietly signing its own supplier contract, and the two need different responses.

There is no universal benchmark for spend per location, because it depends entirely on the type of site and what it does. The useful benchmarks are about distribution and data quality rather than an absolute pounds figure. The questions worth benchmarking are how concentrated spend is, how much spend is unassigned, and how far the normalised cost of each site sits from the median.

Treat these as directional rather than fixed. A new site being fitted out will sit far above the normalised median for a period, and that is expected rather than a problem. The benchmark that matters most is consistency: the same site type, in a similar market, should land in a similar place once you correct for size.

Improving location-based spend analysis means two things at once: making the analysis more reliable, and acting on what it reveals. Cleaner attribution makes the numbers trustworthy, and clear ownership of each location turns an outlier into a fix.

The metric tree approach starts by finding the location with the largest gap between its normalised cost and the median for its type, then drilling into the branch that explains the gap. If the gap is facilities, the lever sits with the property team. If it is local procurement, the lever sits with the site manager and the buying contract.

KPI Tree lets you model this by connecting each location and each cost branch to the team that influences it. Every node carries RACI ownership, so the accountable owner for a site is explicit, and when spend at a location moves outside its expected range the platform pushes that change to the owner rather than waiting for the next review. The verified impact loop then checks whether the renegotiation or consolidation actually moved the number, so you can tell a real saving from a one-off dip.

1. 1

   Attributing to the paying office, not the consuming site

   When a central finance team pays every invoice, naive analysis credits all spend to head office. The location dimension collapses and the whole analysis becomes meaningless. Always map spend to the site that received the goods or service.

2. 2

   Comparing sites on raw totals

   Ranking locations by absolute spend punishes large sites and hides inefficiency at small ones. Without normalising for size, the biggest site always looks like the problem when it may be the most efficient per head.

3. 3

   Ignoring unassigned spend

   A large bucket of untagged transactions quietly distorts every location share. If 15 percent of spend has no location, no ranking built on the other 85 percent can be trusted. Track and reduce the unassigned bucket first.

4. 4

   Mixing one-off and recurring spend

   A site mid-refurbishment will dwarf its peers for a few months. Treating that capital spike as ongoing run rate flags a healthy site as an outlier. Separate one-time project spend from recurring operational spend.

5. 5

   Stopping at the ranking

   A league table of sites is a description, not a decision. Without decomposing each high-spend site into its cost drivers, you know which location to look at but not what to do about it.