BOI™
Bid On Intent.
BOI™ (Bid On Intent) is JudeLuxe's proprietary Google Ads bidding methodology. Every SKU in your catalogue is assigned one of five commercial jobs - Scale, Profit, Protect, Recovery, or Gateway - based on margin, stock, and customer-acquisition role. Bids are set against the job, not blended targets.
By Chris Avery, Co-founder, JudeLuxe · Published 18 May 2026 · Last updated 18 May 2026
Inside the framework
Three letters. One commercial discipline.
Stands for
Bid
Bid against the job, not the blend.
Every bid is set against the single commercial job that SKU is carrying right now. A Profit SKU and a Recovery SKU never share a target, even when they share a campaign. Blended ROAS is what we refuse to optimise to.
Stands for
On
On one job. Never two.
Scale, Profit, Protect, Recovery or Gateway. Each SKU is assigned exactly one job at a time. The job is what the business needs from that SKU this week - not what the platform wants to optimise for.
Stands for
Intent
Commercial intent. Reassigned dynamically.
Intent here means commercial intent: share, margin, defence, cash or downstream value. Inventory shifts, cash position changes, demand moves - and the job a SKU is doing changes with it. The P&L sets the intent. We bid to it.
What BOI™ is not.
What it is
A trademarked operating method for ecommerce Google Ads accounts where contribution margin - not ROAS - is the score that counts.
What clients see
Outcomes BOI™ is built to produce.
- Spend rebalanced from low-intent waste to high-intent demand within the first 30 days.
- Contribution margin per pound of media reported alongside ROAS, every week.
- Performance Max constrained to the SKUs and intent bands where it actually pays.
- A clear answer, every Monday, to the question: did last week's spend pay for itself?
Origin
Why we trademarked it.
BOI™ was born inside accounts where the dashboards said one thing and the bank balance said another. ROAS up, contribution margin down. Conversions up, profit per order collapsing. Performance Max scaling, hero SKUs starving.
After running the same diagnostic across 75+ ecommerce brands, Chris Avery and Gee Kullar codified the pattern into a repeatable method: bid against margin, spend on real demand, segment by intent. We trademarked it so it could not be diluted by the next agency to copy the language.
Trademark record
- Mark
- BOI™
- Stands for
- Bid On Intent
- Owner
- JudeLuxe Ltd
- Authors
- Chris Avery,
Gee Kullar - Applies to
- Google Ads,
Microsoft Ads
The problem BOI™ exists to solve
Why ROAS-led bidding quietly drains profit.
Failure mode
Blended ROAS
Treats a £200 hero SKU and a £15 gateway SKU as equivalent revenue. The blend hides the loss-makers.
Failure mode
Branded conversions
Captured demand reported as paid wins. Spend goes up, incremental sales do not.
Failure mode
Performance Max default
Asset groups optimise to whatever converts cheapest, not to what carries margin. Shopping leakage by design.
Failure mode
Match-type strategy
Treats match type as the lever. Ignores that the same query can be high-intent or low-intent depending on stage.
Failure mode
Last-click attribution
Funds the closer, starves the opener. Intent at the top of the funnel never gets a budget line.
Failure mode
Monthly reporting cadence
By the time waste is visible, four weeks of spend have already cleared. BOI™ runs on a weekly P&L beat.
The five jobs
Every SKU has one job. Never more than one.
At any given week, a SKU is doing exactly one of five jobs. Inventory, cash, margin and demand decide which. The job is reassigned the moment the business needs it to change.
| Job | Business need | Bidding posture | Success metric |
|---|---|---|---|
| Scale | Take share. Fund a launch, a season or a category push where margin can be traded for growth. | Bid up to a defined margin floor. Performance Max allowed off the leash inside guardrails. | Units / new orders at margin floor |
| Profit | Protect contribution margin and cash. Default job for hero SKUs in steady demand. | Bid only where CM3 stays positive at order. Cap branded cannibalisation. | CM3 per order |
| Protect | Defend share of voice on competitor-attacked or range-critical SKUs. Losing visibility here costs more than the bid does. | Hold top-of-page impression share. Margin floor relaxed for defensive queries; tight elsewhere. | Impression share at top vs SOV target |
| Recovery | Convert ageing or end-of-line stock to cash before it becomes a write-down. | Aggressive bids against intent-rich queries. Margin floor lowered, sometimes negative on order, positive on cash recovered. | Cash recovered vs holding cost |
| Gateway | Buy first-time customers or drive attached-basket sales where downstream value justifies a CAC the first order can't. | Bid to allowable CAC or basket-attach value, not order-level ROAS. LTV and attach-rate data feed the next bid. | CAC vs blended LTV / attach value |
One SKU. One job. Reassigned weekly against the P&L, not the platform.
Operating cadence
How BOI™ runs week to week.
Monday
P&L beat
Last week's media spend reconciled against contribution margin. Anything that did not pay for itself gets flagged before the day ends.
Tuesday
Intent re-segmentation
Search term reports rebanded into transactional, comparison, research and retention. Drift since last week is the brief.
Wednesday
Bid resets
Margin floors and intent caps applied. Performance Max asset groups constrained to SKUs that earn the right to scale.
Thursday
Test deployment
One controlled test per account: a new intent band, a feed change, a new audience signal. Always with a margin hypothesis attached.
Friday
Decision log
Every change recorded against expected and actual margin impact. The log is what makes the method auditable, not just operational.
Worked example
ROAS 4.2 looked good. Then we ran BOI™.
A £180k/month spending homewares brand came in reporting blended ROAS of 4.2. The board was happy. Cash was not. Here is what re-banding by intent revealed inside 30 days.
Before BOI™
- Blended ROAS4.2
- Branded share of spend31%
- CM3 per order-£3.40
- Hero SKU spend share14%
After BOI™ · day 30
- Blended ROAS3.6
- Branded share of spend11%
- CM3 per order+£11.80
- Hero SKU spend share38%
Lower headline ROAS. Higher contribution margin. The board now reviews CM3 per order alongside ROAS, every Monday. That is the BOI™ trade.
How it differs
BOI™ vs the standard ecommerce playbook.
| Decision | Typical agency | BOI™ |
|---|---|---|
| Bid signal | Conversion value / ROAS target | Contribution margin per intent band |
| Segmentation | Match type, campaign type | Buyer intent, SKU role |
| Performance Max | Run wide, trust the algorithm | Constrained to earned SKUs only |
| Branded search | Counted as paid wins | Stripped from incremental view |
| Reporting cadence | Monthly platform deck | Weekly P&L beat with decision log |
| Success metric | ROAS, conversions, CPA | CM3 per order and per pound of media |
Worked example
One SKU, one quarter, multiple bid logics.
Illustrative example. A single travel-mug SKU rotates through three jobs across one quarter as stock and business need change. Without BOI™, the same SKU would have run on one blended target all quarter - over-bidding in week 9, under-bidding during the launch window in week 4.
| Week | Stock position | Business need | Assigned job | Bid logic |
|---|---|---|---|---|
| 1–3 | Healthy | Steady profit | Profit | POAS target 2.5x |
| 4 | Healthy | New colour launch | Scale | POAS target 1.5x, capture velocity |
| 5–8 | Healthy | Steady profit | Profit | POAS target 2.5x |
| 9 | Overstocked | Cash flow pressure | Recovery | POAS target 1.1x, max bid |
| 10–12 | Normalised | Steady profit | Profit | POAS target 2.5x |
Implementation
How to implement BOI™ (or a version of it).
- 01
Map your SKU catalogue by margin band
Calculate true contribution margin per SKU: revenue minus COGS, shipping, payment fees, return rate cost, and any promotional discounting. Group SKUs into 3–5 margin bands. This is the foundation. Without it, no bidding logic works.
- 02
Define your five jobs in commercial terms
Scale, Profit, Protect, Recovery, Gateway. Write the rules: what POAS target applies to each, what bid ceiling, what time horizon. Profit jobs run indefinitely. Scale jobs are typically 4–8 week windows. Protect jobs run while the competitive threat or range-criticality remains. Recovery jobs run until stock clears. Gateway jobs require LTV or basket-attach data.
- 03
Set up the operational cadence
Weekly review of every SKU's assigned job. Reassign based on stock, cash, demand. This is not automation. It is judgement applied weekly. The cadence is what makes it work - set-and-forget BOI™ degrades into blended targets within a quarter.
- 04
Feed the signals into Google Ads
Use custom labels in your Merchant Center feed to tag each SKU with its current job. Build Performance Max asset groups by job. Use conversion value rules to feed margin-adjusted POAS targets into Smart Bidding. The platform mechanics matter - but they are the easy part. The hard part is the weekly judgement.
- 05
Report against the job, not the campaign
Build a reporting view that shows POAS and contribution margin by SKU job. "Profit SKUs delivered 2.4x POAS, Recovery SKUs cleared £42k of ageing stock at 1.1x POAS, Scale SKUs grew launch velocity by 38% week-on-week, Protect SKUs held 92% top-of-page IS on defended terms." Compare to typical agency reports that show campaign-level ROAS only and tell you nothing about what is actually happening.
Where it goes wrong
Common mistakes we see when teams try to run BOI™ internally.
Assigning more than one job to a SKU
A SKU cannot be a Profit job and a Scale job at the same time. The bid logic conflicts. Pick one. Reassign as needed.
Letting Scale jobs run indefinitely
Scale is a time-boxed posture. After 4–8 weeks, either the SKU has built the velocity you wanted (graduate to Profit) or it has not (rethink the strategy). Scale jobs that run forever quietly become structural margin leaks.
Skipping the weekly reassignment cadence
BOI™ depends on the cadence. Set it up once, leave it for a quarter, and you have effectively reverted to blended targets with extra steps. The discipline is the moat.
Not feeding LTV or attach data into Gateway jobs
Gateway logic only works if you can value the lifetime customer or the downstream basket, not just the first purchase. Without LTV or attach-rate data, "gateway" bids become "expensive sales" bids.
Questions we get
BOI™, answered straight.
- Is BOI™ software?
- No. BOI™ is a trademarked operating method, not a tool. It runs inside standard Google Ads and Microsoft Ads, supported by our own contribution margin reporting.
- Does it replace Performance Max?
- No. It constrains it. Performance Max keeps running, but only on the SKUs and intent bands where it earns the right to scale.
- How quickly will we see margin shift?
- We typically see CM3 per order move inside 30 days, with the largest gains coming from re-cutting branded share and hero SKU funding.
- Do we need a CFO involved?
- We need access to your COGS, fulfilment cost and refund rate. If those live with finance, finance is in the room. If they live with you, even better.
- Can BOI™ be applied to a low-spend account?
- We apply it from £10k/month media spend up. Below that, the data density is not high enough for the intent bands to behave reliably.
- Who owns the BOI™ trademark?
- JudeLuxe Ltd. The framework was authored by co-founders Chris Avery and Gee Kullar.
Continue reading
Where BOI™ meets the rest of the operating system.
See BOI™ applied to your account.
We'll walk through your current bidding logic, show where intent and margin are being mispriced, and how BOI™ would re-shape it. Minimum £10k/month media spend.
Book a BOI™ review