Real-world application: agency management in practice

Over three days in July 2010, a small group sat in a room at Wieden+Kennedy in Portland with Isaiah Mustafa, a shower set, a towel and a live feed of everything being said about Old Spice on Twitter, Facebook, Reddit and YouTube. They wrote, shot, edited and published something in the order of 180 personalised video replies. Several were online within minutes of the question being asked. No agency reaches that tempo because a statement of work told it to. It reaches it because the client had already settled who could say yes, and that person was sitting in the room.

This lesson stays inside that one relationship: P&G, Old Spice and W+KKThe average number of new users each existing user generates through referrals. Above 1.0, growth compounds on itself and becomes exponential.View full definition →, from the brief written in late 2009 through the response films and the years that followed. Take the roster shapes, scope boundaries and fee mechanics as the foundations lesson sets them out. What matters here is the plumbing underneath a famous campaign: what the brief actually said, where approval sat, how fast the loop closed, and what the speed stopped buying once the format was copied.

Sub-concept 1: a brief with one constraint

The insight P&G handed the agency was narrow and commercial. Household research had shown that a large majority of men's body wash was bought by women, so the brand had been talking to the wrong person in the bathroom. The brief asked W+K to make Old Spice relevant to the woman doing the buying without making the man who uses it feel excluded. That is one audience problem and one non-negotiable. Everything else, tone, casting, medium, was left open, which is why the answer could be a shirtless man on a horse addressing "ladies" directly.

Compare that with the briefs most agencies receive: eight mandatories, three target segmentssegmentsDividing a market into distinct groups of customers who share similar needs, characteristics or behaviours, so each group can be served with a tailored approach.View full definition →, a legal annex and a tonal instruction that contradicts the media plan. A brief with one constraint gives the agency somewhere to be brave. A brief with eight gives it somewhere to hide, because any weak execution can be defended as compliant. The timing sharpened it further: Old Spice was losing ground to Axe, and the launch landed in the same February 2010 window that Dove Men+Care used for its Super Bowl debut. The brand had one shot at owning the conversation in that fortnight and knew it.

Sub-concept 2: approval latency is the real variable

The response week is worth studying as an operations problem rather than a creative one. Roughly 180 films in about three working days is one finished asset every twenty minutes or so of studio time. Run those through a conventional client approval chain, brand manager to marketing director to legal, with a next-morning turnaround, and the ceiling drops to perhaps three or four a day. The creative idea would have been identical. The output would have been two percent of what shipped, and the cultural moment would have closed before the tenth video went up.

Speed of that kind is bought in advance, not in the moment. It needs a named approver with real authority present rather than reachable, a no-go list agreed before the camera rolls (which people, claims and topics are off limits), and a standing kill switch if the tone turns. Once those three exist, the client stops approving assets and starts approving boundaries. Most organisations never measure their own latency. Time the gap between an agency sending work and receiving a decision on it, across ten assets. Two working days is common. It is also an implicit instruction to your agency never to propose anything time-sensitive.

Sub-concept 3: what the numbers did and did not prove

The audience response was not in dispute: tens of millions of views in the first week, the most-viewed sponsored channel on YouTube at the time, Twitter following up by a multiple rather than a percentage, and a Cannes Film Grand Prix plus an Emmy for the campaign. The sales story was messier. Reported figures ranged from a rise of roughly half over three months to a doubling in the month after the response films, and P&G was running price promotion and pushing a repackaged line in the same window.

That variance is the edge case worth carrying away. A campaign that generates this much earned attention almost always runs alongside trade support, because the commercial team sees the wave coming and pushes to meet it, which makes clean attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → impossible after the fact. If you want to know what the creative did, agree the read before launch: a holdout market, a pre-registered window, or a brand tracking measure that promotion cannot flatter. Otherwise you will spend the following year arguing with your finance director about a number neither of you can defend.

Sub-concept 4: what the speed stopped buying

Latitude of the Old Spice sort has a decay curve. In 2011 the brand ran a follow-up with Fabio, including another live response duel, and it did not reproduce the effect. Nothing about the governance had worsened. The format had simply become common property: within about a year, real-time reply videos were a category convention, and the same operating cost bought a fraction of the attention. That is the second-order consequence CMOs underestimate. You are not buying a repeatable mechanic, you are buying a temporary asymmetry, and the asymmetry is the first thing competitors copy.

There is an internal cost too. Once a marketing organisation has seen a war room work, brand teams start requesting "an Old Spice thing" while quietly reinstating the eight mandatories, the committee sign-off and the four-day legal review. The format arrives without the conditions that made it viable. Notably, the relationship itself survived the weaker sequels: W+K has held Old Spice for well over a decade, through several brand platforms. Judging a partner on the year they had a hit is easy; keeping them through the year they did not is where the access actually gets built.

Real-world cases

Case 1: the responses themselves. Mustafa replied to individuals as readily as to celebrities, including a man who used the campaign to propose to his girlfriend on camera. Almost none of that was in the original plan, which was built around broadcast films. The team had permission to redirect a live campaign in the direction the audience was pulling it, without reopening the brief. That permission is the deliverable. Everything else was production.

Case 2: Burger King, as the counter-example. BK ran a similar high-latitude model with Crispin Porter + Bogusky through the 2000s, producing Subservient Chicken and the King. When 3G Capital acquired the company in 2010, the executives who sponsored that latitude left, and the relationship ended the following year. Later work with DAVID shows the split clearly: Whopper Detour in 2018 tied creative freedom to an owned mechanic, a one-cent Whopper unlockable near a McDonald's, and drove the app to the top of the App Store with over a million downloads during the promotion. Moldy Whopper in 2020 won heavily at awards shows and is still argued about commercially. Same latitude, different amount of business machinery attached to it.

CMO action items

• Measure your approval latency on the next ten pieces of creative. Timestamp when the agency sends and when a decision lands. If the median is over 24 hours, nothing time-sensitive will ever be proposed to you.
• Take your next brief and cut it to one audience insight and one non-negotiable. Everything you delete, write down separately as a judgement you are delegating, and name who holds it.
• Before any live or reactive activity, agree the no-go list, the single named approver in the room and the kill trigger. Do it in writing, once, rather than per asset.
• Agree the measurement read before launch, including which promotional activity will run in the same window. Decide now what result would make you stop.

Common mistakes that kill results

Mistake 1: approval by committee. When legal, brand, regional leads and the CMO all hold a veto, agencies stop proposing anything with risk in it, because the expected cost of a bold idea dying in review exceeds the expected reward. Pick one decision-maker per campaign type and defend that choice publicly the first time someone senior is bypassed. That first test is the whole thing.

Mistake 2: importing the format and leaving the conditions behind. The Old Spice war room worked because of a one-constraint brief, pre-cleared boundaries and an approver with authority in the room. Ask for the output without rebuilding those, and you get an expensive studio day producing content nobody can sign off.

Mistake 3: giving feedback on executions without reconfirming the brief. "Make the logo bigger" and "I don't like the palette" solve the wrong problem. The only useful judgement on an execution is whether it delivers what was asked. If the brief was wrong, say the brief was wrong and take that cost yourself. Aesthetic preference dressed as strategy is the fastest way to teach an agency to stop showing you its first idea.