Founders ask the same fork in three outfits: keep the agency, hire a buyer, or buy software. The wrong question is “which is best?” The right question is “what is broken: taste, overnight ops, creative supply, or offer?”
Side-by-side without the pitch deck
| Agency retainer | In-house buyer | Operator software | |
|---|---|---|---|
| Cost shape | Monthly multi-thousand typical | Salary + tools + management | Software pricing; you keep the keys |
| Best at | Strategy, specialised craft, surge capacity | Day-to-day account craft and speed | Stock/ads correlation, briefs, review queues |
| Weak at (often) | Overnight stock discipline if not scoped | Burnout when they are also the founder | Brand taste without a human |
| Stock mid-flight | Only if written into the retainer | If they open Shopify daily | Designed as a first-class input |
| Failure mode | Decks and junior hours | Heroics until they quit | Chat decoration with no actions |
Thicker retainer unpack: Shopify retainers. Buyer decision: when a media buyer is worth it. Money model: agency cost tool.
Decision rules that stay honest
- Bottleneck is taste / brand narrative: humans first (agency or senior hire).
- Bottleneck is overnight monitoring, stock/ads mismatch, packaging for approval: software desk first.
- Bottleneck is pure auction craft at high spend: specialist buyer or performance pod.
- Bottleneck is weak offer: none of the three will save you. Fix the offer.
Hybrids that work
Agency on creative strategy, software on daily ops. Or in-house buyer plus software for stock truth and morning briefs. The broken hybrid is “everyone owns everything,” which means nobody owns the Friday OOS.
Ralph is built for the software side of that hybrid: Shopify-centred operator desk, Meta/Google context, stock intelligence, campaign packages in review. Pillar: Ralph vs agency. Category frame: all-in-one ecommerce AI.
Pay for decisions and assets. Refuse to pay for commentary dressed as strategy.
90-day review questions
- Did stockouts still surprise paid?
- Did we ship assets or only commentary?
- Is the founder still the night-shift buyer?
- Can we name who pauses a dead size on a Friday?
If answers are soft, change the shape. Do not renew on vibe.
Questions people actually ask
Should I fire my agency for software?
Only if they are selling babysitting and decks. Keep sharp humans for taste and strategy; replace overnight legwork with systems.
When is an in-house buyer worth it?
When creative supply and spend volume justify a full-time craft role. Not as a substitute for a clear offer.
Can software replace a media buyer?
It can shrink monitoring and packaging. It should not pretend to own brand risk alone.
What is the hybrid that usually works?
Human taste + software ops desk + optional specialist agency lanes (creative or channel craft).
If the retainer is the problem: Ralph does the routine half of a retainer every night from £99 a month: watching stock against spend and drafting the next package. Keep your humans for taste and strategy. See Ralph · What an agency costs · Docs