The machine, not the entry
Breaking down our painfully learned structural approach to awards

The first year we took awards seriously at Mindshare China, in 2014, we had no process at all. Someone remembered in September that entries were due, a folder got opened, and whoever was around that week wrote the drafts. We were making case study videos with surfer dudes in Bali who did video production on the side and occasionally just vanished. We won a gold, a silver, and two bronze. That somehow turned out to be enough to win Agency of the Year for Smarties China.
By 2020 the same program had produced more than 1,000 entries and 300 wins, then a run of top-10 finishes in the WARC Media 100 and a Global Media Agency of the Year silver at Campaign in 2023. None of that came from writing better entries than everyone else. It came from building a machine that constantly learned from its mistakes, evolving with every cycle, so the quality did not depend on who happened to be free that week.
The machine was five stations. Here is what each one actually did, and why we ran them in this order.
1. Collection and screening
This is the least glamorous station and the one that saved the most money. Every cycle, the suggestions come in. Some are from people who understand what wins and most are not. You get campaigns that did real numbers but have no story, campaigns with a great story and no numbers, and a lot of "we should enter this" with nothing behind it but that someone liked the work.
The job here is not to draft anything. It is to go find the actual results, chase down the data that may or may not exist, and form an early read on which handful of these has something real in it. Most do not. Saying so at this stage, before anyone has spent a fee or a weekend, is the entire point of the station.
Do this badly and you carry ten weak campaigns all the way to the drafting stage before anyone admits most were never going to place. Do it well and you are working on three or four that have a genuine shot.
2. First drafts
The word "draft" is misleading here, because this station is not about writing full entries. It is about angles.
Take the campaigns that survived screening and work the options. What is the actual claim. Which show, which category. Is this a media story, a creative story, a data story. What is the hook that makes a tired juror look up. We would spin the same campaign three or four different ways and see which framing had something in it and which fell apart the moment you said it out loud.
Most of these angles die here, and that is the station working. You are not trying to finish anything. You are trying to find the version of the case worth finishing.
3. Review cycle
This is where the angles get validated or killed by people who have actually sat on juries.
At the network we ran it as local, regional, and global councils. A senior person, usually someone who had judged, would read what we had and tell us where the potential was and where we were kidding ourselves. Someone in New York would look at a campaign from China and say something that frustrated the hell out of us, and we would have to take it into account anyway, because that is who ends up in the room at the global shows.
The value of this station is entirely in who runs it. Anyone can list ten things wrong with a draft. Someone who has judged can tell you which one actually costs you the medal, and where a case you had written off still has room. That judgment is the rarest thing in the whole process, and this is the station that applies it.
4. Script generation
This is the one that surprises people. In most cases we drafted the two-minute video script before we wrote the full entry.
The logic is simple. The script forces the case down to two minutes: the context, the insight, the idea, why it worked, the result. If the campaign cannot survive that compression, if you cannot explain it coherently to a jury in the time they will actually give it, the full entry is not going to save it. So we used the script as the test. Does the insight travel outside its home market. Do the mechanics make sense to someone who was not there. Do the results land quickly and clearly, or do they need a paragraph of setup nobody will read.
Writing the script first tells you whether the case holds before you invest in the long-form entry. If the two-minute version does not work, you go back to station two and find a different angle, or you drop the campaign. Better to learn that now than after the entry is written and the fee is paid.
5. Final draft edits
Only now does the full entry get written and polished, and only for the campaigns that made it through the first four stations intact.
This is where the work gets specific to the category and the jury. The same campaign gets edited differently for a media category than for a creative one, differently for a jury in Asia than for a global panel. This is the polish stage, the tuning, the "does this exact word land with this exact room" pass. It is real work and it matters, but it is the last ten percent, not the first ninety. An entry that reaches this station has already been screened, angled, validated, and pressure-tested against a two-minute cut. The polish is finishing a case that was already sound, not rescuing one that was not.
Most agencies spend all their time here and almost none on the four stations before it. That is exactly backward, and it is why so much award effort goes into polishing entries that were never going to place.
Why the order is the whole point
Every station kills work the next one would otherwise waste time on. Screening kills the campaigns with no story before anyone drafts an angle. The angle stage kills the framings that fall apart before anyone builds a case. The review cycle kills the cases a juror would reject before you script them. The script kills the entries that cannot survive two minutes before you write the long version. By the time something reaches final edits, it has earned the polish.
Run it out of order, or skip stations, and you get what most programs get: a scramble in September, real money spent on entries that died at the first station nobody ran, and your best people doing all of it on weekends and holidays because the process only exists during deadline week. I spent more holidays than I can count editing entries that screening should have killed in an afternoon.
The other thing the order buys you is memory. Run the same stations, with the same people, cycle after cycle, and the machine gets smarter: screening starts catching the weak campaign faster, the review cycle remembers which categories were traps last year, the whole thing compounds because it kept a record of what placed and what didn't. How to build that feedback loop deliberately, across two cycles, is a discipline of its own, and its own piece in this run.
The independent-agency version
None of this needs network headcount. A 20 to 50 person shop runs the same five stations at a smaller size. Screening is one afternoon going through the slate before anyone drafts. Angles is one person spinning the survivors a few different ways. The review cycle is one colleague who has judged, reading the angles before the fees are paid. The script test is writing the two-minute version first, on any campaign, to see if it holds. Final edits is the polish, last, only on what survived.
The stations do not need to be elaborate. They need to happen in order, every cycle, instead of collapsing into one adrenaline week because nobody scheduled the other four.
Shortlist exists to run the early stations without a dedicated award manager: mapping the slate against real show and category fit, pressure-testing the angle, drafting and scoring the two-minute script against a specific show's criteria, all before the full entry gets written. It does by default what the first four stations do by hand. Running them yourself, on a fixed schedule, works just as well. The machine is the point, not the tool that happens to run part of it.
Your final edit, brought forward as fast as possible.
30+ shows. Unlimited entries. The judgment stays yours.
Start free trial