The Banner That Didn't Fit the Shop

The starting condition

A partnership only becomes visible to a customer if something in the physical location says so. Dokitami's field officers could sign a pharmacy, a gym, or a supermarket in a single conversation, but signing a business and making that business actually look like a Dokitami partner to the people walking through its door were two different jobs, and the second one depended entirely on physical materials arriving, fitting, and staying up.

The department settled early on a standard: a 27 by 40 inch banner, printed on paper rather than the more expensive vinyl roll-up, as the default way to equip a newly signed location. It was cheap once produced at volume, cheap enough to hand out at scale to a fast-growing partner list. Alongside it came flyers, produced in large batches. By the middle of the campaign, banners and flyers had gone out across five distribution hubs in Alimosho, Surulere, Ikeja, Apapa, and Ikorodu.

The size and format had been picked before more than a handful of stores had actually tried to display anything. It was a reasonable default: big enough to be seen, cheap enough to print at volume, standard enough that officers didn't need to make a judgment call in the field about which material to carry. It was also, for a meaningful share of the partner base, close to unusable.

The question that mattered

The easy framing, once field reports started coming back with complaints about materials, is "we need to print faster" or "we need to deliver materials sooner." That framing assumes the format itself was right and the problem was logistics. The harder and more useful question was whether one material, in one size, could actually equip every kind of location the department was signing — a small pharmacy counter, a gym reception wall, a supermarket entrance, and a POS agent with no shop at all, just a table and an umbrella on a street corner. Those are not the same physical space, and treating them as though a single banner would work in all of them was an assumption nobody had tested before committing to volume.

What was tried, and what the evidence showed

The printer relationship itself was the first place the department had to solve a real, unglamorous problem: getting materials made at all, on acceptable terms and on a schedule that didn't leave the field waiting. That negotiation mattered, because every day of lead time on materials was a day a signed partner sat unequipped and, functionally, inactive.

Once material started reaching the field, the reports made the size problem visible fast. Officers reported banners too wide to fit inside small shops, banners that blocked a doorway or a till rather than sitting cleanly on a wall, and outlets with simply no safe or sensible place to mount something that size. Some partners asked for something smaller, a tabletop stand, an A2-sized version, a sticker, because the standard banner was more disruptive to their space than it was helpful to Dokitami's visibility. POS agents were the clearest mismatch of all: the whole appeal of testing that channel had been that a per-referral payout mattered more to an individual agent working thin margins than it did to a shop owner, but a wall banner is meaningless to someone who doesn't have a wall. The format that worked for a pharmacy counter did nothing for the exact channel the department had identified, weeks earlier, as one of its most promising and least explored.

There was a second, quieter problem layered on top of the size issue: attribution. Several outlets objected specifically to the phone number printed on the banner, worried that a customer who called that number directly wouldn't be recognisably linked back to their store, and so wouldn't earn them anything. A materials decision that looked purely like branding — what goes on the banner — turned out to also be a commercial-control decision about whose name a referral belonged to, and it wasn't resolved cleanly before the archive ends.

The system built

The fix that emerged wasn't a redesign of the banner. It was a rule about matching material to location type, issued once enough field evidence had accumulated to make the mismatch undeniable. The 27 by 40 banner was explicitly designated for store and business partners with a physical wall to use it on — not for POS or individual operators, who needed something else entirely. Officers were directed to use their portal ID at the point of collection and to provide photo evidence once a material was actually displayed, which turned "materials delivered" from an assumption into something that could be checked against a photograph rather than taken on trust.

The underlying principle, stated plainly once it had been learned rather than assumed: equipping a partner is a fit problem. The fit depends on the physical shape of the location, not on a single company-wide standard. A pharmacy counter, a gym reception desk, and a street-corner POS table each needed their own answer — a wall banner for the first, something smaller or nothing at all for the second two — and the department's job was to figure out which locations needed which format rather than shipping the same material everywhere and hoping it fit.

What changed, honestly stated

By the time detailed field evidence had accumulated, the department could point to real, if uneven, activation: reports from single officers showing figures like 30 activations out of 36 signed-store revisits, or 25 equipped stores from a batch of follow-ups. Those are self-reported, single-officer counts rather than an independently verified network-wide total, and they should be read that way — as evidence that the corrected approach was producing activation, not as a precise account of how many of the roughly 1,500 signed partners were genuinely equipped by the end of the period covered here.

What is clear, and doesn't need the precise total to be true, is that a meaningful batch of banners had already been printed and distributed to one specification before the size guidance changed — a real, dated cost of having made a materials decision ahead of the field data that would have informed it better. The POS-agent channel, identified in the department's own first field test as one of its most promising, was still without a workable physical material of its own by the end of the record: an open thread, not a solved one.

What this generalises to

A material or a process that is standardised before it has been tested against the actual variety of conditions it will meet is optimised for the average case, and very few real cases are the average case. A pharmacy counter and a POS agent's table are not variations on the same problem; they are two different problems that happen to share a company name. The cheaper path — pick one format, print it in bulk, distribute it everywhere — looks efficient right up until the field reports come back, and by then the volume has already been committed.

The more durable habit, visible once the department corrected course, is treating "equipped" as a claim that needs evidence — a photograph, a specific material matched to a specific kind of location — rather than a status a partner is assigned the moment something is dropped off. A banner sitting in a stockroom because it didn't fit the wall counts as delivered in a spreadsheet and counts as nothing in a customer's eye line. The gap between those two facts is exactly where a distribution channel quietly stops working without anyone noticing until someone asks for a photograph.