Every multi-location business runs the same argument on a loop. National marketing wants scale, consistency, and one set of numbers. The locations want their own market's truth: this store's competitors, this city's pricing, this manager's Saturday event. Co-op dollars sit between them like a disputed border, and the media plan ends up shaped by whoever won the last meeting rather than by how customers actually buy. Automotive runs this war in its purest form, but franchises, banks, gyms, and medical groups all fight the same one, and I have refereed it in most of those categories.
The mistake is treating it as a contest with a winner. It is a layering problem. National money is unbeatable at what only scale can do: brand demand, broad reach, negotiating leverage, creative production nobody's local budget could fund. Local money is unbeatable at what only proximity can do: the seven-mile radius where the store actually competes, inventory and staffing truth, the reviews and maps presence that decide which location a ready buyer visits. Each layer performing the other's job is where the waste lives. National buying zip-code minutiae is expensive dabbling. Locations buying brand awareness is a rounding error announcing itself.
The customer does not know your org chart
Watch the purchase journey and the layering stops being theory. A customer sees national brand advertising for months, forms a preference she could not date or attribute, then one Tuesday searches the category, checks a map, reads four reviews of the two nearest locations, and picks one. National built the demand. Local captured or fumbled it, usually on assets that cost almost nothing: the Google Business Profile with current hours, the review response from an actual human, the location page that loads and says whether the thing is in stock. I have seen seven-figure national campaigns lose their conversion at a maps listing showing the wrong closing time. The customer experienced one brand. The brand experienced two departments not speaking.
That journey dictates the architecture, and the architecture that works is boring and rare. National owns the demand pool and the standards: brand media, creative systems, the measurement spine. Local owns capture inside a defined fence, with templates that keep the brand intact and enough freedom to be true about price, stock, and people. Co-op funds route through a shared platform so spend is visible in one place, because the alternative, hundreds of separately managed local accounts, is how fees and duplication quietly eat the budget. When I audit multi-location media, duplicated management fees across local accounts are reliably the first six figures of savings found, before a single targeting decision improves.
Measure the seam
Here is the discipline almost no multi-location marketing organization runs, and it is the entire diagnostic: measure the handoff, not just the layers. Every market has two numbers, the demand national created there, branded search volume is a fine proxy, and the share of that demand the local layer converted. Put them on one grid, markets down the side. Four quadrants fall out. High demand, high capture: leave it alone. Low demand, high capture: national should spend more here, the local operation has earned it. High demand, low capture: the expensive quadrant, national dollars pouring into a leaky bucket, and the fix is local, cheap, and fast. Low demand, low capture: a real conversation about the market itself.
The number that exposes a broken system is a market where national demand rose and local capture did not, because it means the organization paid twice, once to create the customer and once to lose her. That is the meeting worth having, and in most companies it has never been held even once, because demand data lives in the national dashboard and capture data lives in the locations, and no one owns the seam.
The budget fight will renew itself every planning season regardless; it is a permanent feature of the structure. But the companies that grow are not the ones that settle it. They are the ones that made it irrelevant by giving each layer the job only it can do, and then watching the handoff like it is the whole business. Because at the moment of purchase, it is.