Hook-first editing
Skippable in-stream lives or dies in five seconds. We cut for the hook, then produce three or four hook variants from the same footage so there is always a test running.
- In-stream
- Hook variants
- 5s discipline
Buy one, buy all six. The reason clients usually end up taking most of them is that the value sits in the joins — the editor sitting next to the person managing the campaign, seeing which asset is winning on the day it happens.
Google has automated most of the levers that used to define account management. Bidding is automated. Targeting is largely automated. Placement is automated. What remains genuinely under your control is what you put into the account and how quickly you replace what stops working.
So we manage accounts around the asset pipeline rather than around a bidding calendar — while still doing the unglamorous work properly: negatives, exclusions, brand separation and conversion hygiene.
Two shoot spaces on Deansgate, a full lighting and camera inventory, and five editors. That is why we can promise a first cut in 72 hours instead of hoping a freelancer is free.
Skippable in-stream lives or dies in five seconds. We cut for the hook, then produce three or four hook variants from the same footage so there is always a test running.
Purpose-built 9:16 edits with safe-area awareness, burned-in captions and text sized for a phone held at arm's length — not a landscape cut with bars added.
Six and fifteen second versions produced from the same session, so frequency and reminder duty is covered without commissioning another shoot.
Every winning concept gets templated so it can produce twenty credible variants without twenty briefs. This is how a modest production budget behaves like a large one.
Open almost any underperforming Performance Max campaign and you will find five stock images, no video, and headlines written in a rush during setup eighteen months ago. Google then auto-generates a video from the images, and everyone concludes the campaign type does not work.
We build complete asset groups per product line, every ratio covered, real video included, and refresh them monthly against asset-level performance data. It is unglamorous and it is usually the single biggest available improvement.
Demand Gen places you in Discover, Gmail and the YouTube feeds — environments where people are browsing content, not shopping. Creative that looks like a display banner is ignored there, politely and completely.
Image sets designed to sit convincingly in a feed: real photography, minimal text overlay, and a proposition that reads before anybody notices it is advertising.
Short vertical video cut for silent autoplay, with captions and a first frame that earns the second one. Produced from the same shoot days as your YouTube assets.
Distinct asset sets per audience segment rather than one creative shown to everybody, because a returning customer and a cold prospect need different arguments.
Kept in separate campaigns with brand excluded, so Demand Gen has to earn its budget on genuine prospecting instead of harvesting existing intent.
They are rarely treated as one. The typical account has fifteen headlines that are variations on the same sentence, keyword insertion doing the heavy lifting, and nobody looking at which asset Google actually preferred.
We write genuinely distinct headlines expressing different propositions, pin only where the brand or a legal requirement demands it, and read asset performance labels monthly to decide what to replace.
Asset IDs tell you nothing you can act on. Grouping assets by the concept they express tells you which argument persuades people — which is knowledge you can use in the next shoot, the next landing page and the next email.
Every asset tagged to a concept and a hook before it goes live, so reporting rolls up to ideas rather than filenames. Eight concepts tested a month on a standard retainer.
Campaign experiments where volume supports them, with a stated hypothesis and duration agreed in advance so nobody calls a winner after four days.
Frequency and performance decay tracked per concept, so assets are retired on evidence rather than when someone internally gets bored of seeing them.
The report opens with what failed and what we are changing. It makes the meeting shorter and the relationship considerably longer.
Most clients take a monthly retainer combining account management and production. We also run standalone projects — a Performance Max asset overhaul, a YouTube campaign launch — and content days where we shoot a quarter's worth of Google Ads assets in one go.
Tell us what you need