WhentoRunGoogleAdsataDeliberateLoss

The default position in paid media management is that every campaign should be profitable on its own terms. ROAS above break-even, contribution margin positive, CPA inside the unit economics. Anything else is a failure of optimisation, an account that needs work, or a client that needs to be retrained. This position is wrong in a specific and important set of circumstances.

There are situations where the correct commercial decision is to deliberately run Google Ads at a loss (not as a mistake, not as a learning-phase tolerance, but as a planned commercial action where the loss on the ad account is the cheapest way to produce a larger gain elsewhere in the business).

Most agencies won't tell their clients this. Some don't know it. Most that do know it won't say it, because it sounds like an excuse for poor performance and the conversation is hard to have without explicit alignment from the founder and the finance team. So the default position holds. Accounts get optimised to "always profitable" targets that produce locally rational decisions and globally suboptimal outcomes.

This article sets out the cases where deliberate-loss advertising is the right commercial call, how to size the loss properly, how to structure the campaign so the loss is bounded and visible, and how to report on it so the finance team sees what the marketing team is doing and why.

What "Deliberate Loss" Actually Means

Before the cases, a definition. Deliberate-loss advertising is a campaign or campaign segment run with an explicit, pre-agreed expectation of negative contribution margin on the directly attributable revenue. The loss is bounded, a maximum budget, a maximum duration, a maximum unit-level subsidy. The loss is justified, there is a separate commercial gain that the loss produces, and the gain exceeds the loss. The loss is reported, finance and operations see the loss in the management accounts, understand why it is there, and have signed off on the framework.

It is not:

• A campaign that is accidentally losing money because the targets are wrong
• A campaign that is losing money because the bidding algorithm hasn't learned yet
• A campaign that is losing money because the agency hasn't audited it recently
• A campaign that is losing money because the attribution is broken
• A campaign that is losing money because the margin assumption was wrong

Case 1: Working Capital Recovery on Aged Stock

The clearest and most common case is the recovery of working capital trapped in aged or off-season stock.

The premise: inventory in an ecommerce business is a depreciating asset. Stock that doesn't move at full price will be marked down, written off, or disposed of. Each step destroys more economic value than the previous one. By the time a unit reaches write-off, the recoverable value is often a single-digit percentage of the original landed cost.

A specific example. A fashion brand holds £200,000 (at retail) of SS25 stock in mid-November. The landed cost was £80,000. Current sell-through at full price has stalled, the season has ended commercially even though the calendar disagrees. The alternatives:

The ad account looks bad in isolation during this period. Blended ROAS drops. POAS goes negative on the clearance segment. The campaign reports show a campaign that is "losing money." But the working capital recovery, £80,000 to £100,000 of capital released back into the business, is significantly larger than the paid media loss, which might be £5,000 to £10,000 of negative contribution.

This is the single most common case for deliberate-loss advertising and the one most agencies handle worst. The reason is that the loss is local to the ad account and the gain is distributed across the balance sheet. If your agency can't tell this story to your finance team, they're costing you money by not running this campaign.

Case 2: New Customer Acquisition With Strong LTV

The second case is acquisition at a deliberate first-order loss when the customer's lifetime value justifies it.

The premise: in subscription businesses, repeat-purchase brands, and high-frequency consumables, the value of a customer is materially larger than the value of the first order. If the lifetime contribution is large enough and the payback period is acceptable, you should be willing to acquire that customer at a first-order loss, sometimes a substantial one.

This isn't novel. The DTC playbook has been built on this principle for fifteen years. What's new, and what most agencies still get wrong, is the structural separation between new-customer and returning-customer paid traffic, and the bid strategy implications.

If your account doesn't separate new customers from returning customers in its conversion structure, the algorithm cannot optimise for them differently. The "blended" bid produces a bid that's too low for new customers (because the returning-customer profitability drags the average down to a target the new-customer segment can't meet) and too high for returning customers (because the new-customer ceiling drags the average up).

The structural fix:

• Separate conversion actions. Configure new-customer purchases and returning-customer purchases as distinct conversion actions in Google Ads. This requires either a customer-status flag passed to the conversion endpoint or a Google Analytics 4 audience-based conversion setup. Both are achievable; neither is the default in most accounts.
• Separate campaigns or bid strategies. New-customer campaigns operate with a ROAS or CPA target derived from the unit economics of a new customer over the LTV window, typically allowing significantly more aggressive bidding than a blended target would permit. Returning-customer campaigns operate with a tighter target reflecting the higher conversion rate and lower acquisition cost.
• LTV-anchored target setting. The new-customer target is not "break-even on first order." It is "break-even at the end of the payback window we've agreed with finance." If the payback window is 90 days and the average new customer makes 2.1 orders in their first 90 days at a contribution margin of £18 per order, the new-customer target should be set against £37.80 of expected first-90-day contribution, not against the £18 contribution of the first order.

This case is well understood in subscription DTC. It is less well understood, and less well executed, in non-subscription DTC where the LTV is real but slower-developing. Coffee, skincare, supplements, pet food, household essentials, replacement cycles in fashion and accessories. All of these have meaningful repeat patterns and most of them are being advertised against blended targets that under-invest in new customer acquisition.

The deliberate-loss framing matters because the first-order economics genuinely are negative under the new-customer target. The agency report needs to show that loss honestly and contextualise it against the LTV recovery. If it just shows a first-order ROAS hitting target, the framing is lost and the discipline degrades.

Case 3: Market Entry and Category Establishment

The third case is structural rather than tactical. When entering a new market or establishing a new category, the early advertising spend is producing search volume, brand recognition, and category understanding that doesn't show up as attributable conversion value for months.

The premise: in a new market, demand is partly latent and partly created. The advertising you run in months 1 to 3 generates search queries, brand searches, and direct visits in months 4 to 9 that the attribution system doesn't connect to the original ad spend. The reported ROAS during the establishment phase is artificially low because half the value being created hasn't materialised yet, and most of what does materialise gets attributed to brand or organic.

Four requirements for legitimate market-entry deliberate loss:

• A defined market or category. Specific geographic market, specific product category, specific customer segment. Not "growth", a defined target with measurable boundaries.
• A defined duration. Typically 3 to 6 months of establishment spend, with a clear point at which the deliberate-loss treatment ends and standard performance targets resume.
• A defined budget. A capped total spend across the establishment period. Not an open commitment, not a percentage of revenue, a hard number.
• A defined success metric beyond ROAS. Branded search volume in the target market, direct traffic to the target category, repeat purchase rate from the cohort acquired during the establishment phase. Something that captures the latent value being created.

If all four are in place, market-entry deliberate-loss advertising is a defensible commercial action. If any are missing, it's brand spend dressed up as performance, and the loss is real but the gain is speculative.

The case is most legitimate for businesses with a track record of successful establishment in other markets, UK brands entering the US, US brands entering the EU, established categories expanding into adjacent ones. The historical pattern provides the calibration: if past market entries have produced repeatable post-establishment performance, the current entry can be modelled against that benchmark. Without historical evidence, the case is weaker, and the loss should be smaller.

Case 4: Auction Defence and Competitor Exclusion

The fourth case is the most contentious. Defensive bidding on branded terms, category terms, or competitor terms where the unit economics of the click are negative but the strategic value of denying the competitor the impression is significant.

The premise: in concentrated competitive categories, the cost of letting a competitor occupy the top of the SERP for your branded queries or your highest-intent category queries is larger than the cost of the defensive clicks. Branded search defence is the cleanest version of this, you may pay £0.40 per click to defend your own brand from competitor intrusion, knowing that the marginal incremental revenue is small but the prevention of intrusion is structurally valuable.

This case is more often abused than the others. Two reasons.

First, branded search defence is frequently sold to clients as profitable when it isn't. Branded search converts well, the ROAS looks excellent, the agency books a win. The incremental analysis (what would happen if you turned the branded campaign off?) almost always shows that the bulk of that revenue would arrive anyway through organic. The reported ROAS is mostly attributing organic-eligible revenue to the paid campaign. The campaign is real, the work is real, but the framing is wrong: most of what looks like profit is cannibalisation.

Second, competitor-conquesting campaigns frequently lose money in absolute terms (low conversion rates, high CPCs, weak intent) but get framed as strategic. They might be strategic for a fraction of accounts. For most, they are a budget line that produces neither attributable profit nor measurable strategic gain.

The legitimate version requires:

• Incrementality testing. Geo-holdouts, scheduled pauses, or true incrementality tests that quantify the actual lift produced by the defensive or conquesting campaign. If the campaign produces no incremental revenue when run versus paused, it isn't defending anything, it's just spending money on traffic that would have arrived anyway.
• A quantified strategic gain. Branded SERP share, competitor click-share, share of voice in target queries, a metric that captures the structural value being defended. Without this, "strategic" is a placeholder for "we can't show ROI."
• An exit threshold. A point at which the defensive activity is reviewed and either scaled, reduced, or stopped. Open-ended defensive spending tends to grow indefinitely because nobody owns the kill switch.

When these conditions are met, deliberate-loss defensive bidding is a legitimate commercial action. When they aren't, it is the most expensive habit in paid media management, a budget line that compounds for years because nobody is willing to test what happens if it stops.

Sizing the Loss Properly

The deliberate-loss decision is only as good as the sizing. Three principles.

A simple loss-sizing framework for clearance:

For new-customer acquisition:

For market entry:

These aren't precise formulas; they're decision frameworks. The point is that the loss size has a defensible derivation, not a number plucked from instinct.

Structuring the Campaign

Deliberate-loss campaigns need structural separation from the rest of the account. Three reasons.

First, the bidding algorithm needs a clean signal. If the deliberate-loss segment is mixed with profitable segments under a single campaign, the algorithm will average the economics and produce a bid strategy that fits neither. The clearance campaign needs to learn that aggressive bidding on aged stock is acceptable; the profitable campaign needs to continue learning that disciplined bidding on hero products is required. Separation is what allows both signals to coexist.

Second, the reporting needs to isolate the loss. The conversation with finance is "we ran £8,000 of loss to recover £85,000 of working capital, here's the campaign, here are the numbers." That conversation is impossible if the loss is embedded in a campaign that also contains profitable activity. The structural separation makes the financial reporting possible.

Third, the kill conditions need to be enforceable. A clearly bounded campaign can be paused, scaled, or terminated without affecting the rest of the account. A loss segment buried inside a broader campaign cannot.

The mechanical structure:

• Dedicated campaign per deliberate-loss purpose. A clearance campaign for aged stock. A new-customer acquisition campaign for high-LTV cohorts. A market-entry campaign for the target geography. Each runs separately and is reported separately.
• Custom labels in the feed to drive segmentation. Products designated for clearance get a `clearance` flag in a custom label. The clearance campaign filters on that label. When the campaign ends, the labels are removed and the products return to the standard structure (or are excluded entirely if write-off is the next step).
• Bid strategies aligned to the purpose. Clearance campaigns typically use Maximise Conversion Value without a ROAS target during active windows, accepting that the algorithm will spend the full budget against the available demand. New-customer campaigns use Target CPA or Target ROAS set against the LTV-adjusted economics. Market-entry campaigns often use Maximise Conversions with a CPC ceiling, prioritising volume of category exposure over conversion efficiency.
• Daily budget caps. Every deliberate-loss campaign has a hard daily budget that, multiplied by the campaign duration, equals or is below the pre-agreed loss cap. The budget is the structural enforcement of the loss limit.

Reporting Deliberate-Loss Activity

The reporting is where most agencies fail at this. The campaign exists, the loss is real, the gain is real, and the monthly report shows a blended ROAS that includes the loss and confuses everyone reading it.

The fix is two-tier reporting.

The two tiers are summed for total account spend reporting (finance still needs the total), but the performance interpretation is kept separate. The headline conversation with the operator is: "Tier 1 is performing at target; Tier 2 is running the clearance program agreed in October and is on track to recover £85,000 of working capital by mid-December against £8,000 of allocated loss."

That conversation is impossible without structural separation in the account and disciplined separation in the reporting. Most agencies don't do either. The result is that deliberate-loss activity, when it happens at all, gets framed as a problem rather than a planned commercial action, and the discipline that makes it work degrades over time.

The Finance Team Conversation

Deliberate-loss advertising requires finance team alignment. Not "informed after the fact" alignment, explicit pre-commitment alignment, ideally in writing.

The conversation has three parts.

Once these three parts are agreed, the activity runs against the agreement. Finance sees the loss in management accounts, knows what it's for, and can stop worrying about it on the marketing line, because it isn't a marketing decision anymore, it's an operational commitment with marketing as the execution channel.

This is the single largest barrier to deliberate-loss activity in most ecommerce businesses. Not the technique. The organisational conversation. Marketing teams don't bring this to finance because they don't have the language for it. Finance teams don't ask for it because they don't know it's available. Agencies don't propose it because the conversation is hard and the default position, "all campaigns should be profitable," is easier to defend.

Getting past this barrier is what separates operators that use paid media as a commercial lever from those that use it as a marketing channel. The first treat the ad account as a tool that can do many things, some of them deliberately unprofitable in exchange for larger gains elsewhere. The second treat the ad account as a profit centre that must always be profitable on its own terms, which sounds disciplined and is in fact a constraint that costs the business money in every case described above.

Where Most Accounts Get This Wrong

In the field, the most common failure modes for deliberate-loss advertising are:

All of these are avoidable. The structural and operational discipline required isn't complex. The organisational discipline required is harder, the willingness to have a specific conversation with finance, the willingness to bound the activity in advance, the willingness to report it honestly even when the numbers in isolation look bad.

The Strategic Position

Most paid media management operates under the implicit assumption that the ad account is a profit centre to be optimised against its own internal metrics. That assumption is true in normal operating conditions and false in the specific cases described above. The operators that recognise the difference run their paid media as a flexible commercial instrument that can produce profit, recover capital, acquire customers, or establish markets, selecting the right objective for the conditions and accepting the right local outcome to produce the right global outcome.

The operators that don't recognise the difference run their paid media as a permanently profitable channel that's structurally unable to address the situations where the right answer is a deliberate loss. They miss aged stock recovery opportunities. They under-invest in new customer acquisition. They fail at market entry. They run defensive bidding that doesn't defend anything.

The cost of the second approach is invisible because the ad account looks healthy. The opportunity cost shows up in the rest of the business, working capital that stays trapped, customer cohorts that never compound, markets that get conceded to competitors, brand SERPs that get colonised by aggressive challengers.

If your agency has never proposed running an activity at a deliberate loss, ask them why. There are three possible answers.

• They have, and you said no. Fair enough; the answer might still be no, but at least the conversation happened.
• They haven't because none of the conditions described above currently apply to your business. Also fair. The technique isn't universally applicable, and an agency that doesn't propose it when it isn't needed is doing its job.
• They haven't because the conversation is uncomfortable, the framework is unfamiliar, or the default position of "always profitable" is easier to deliver than the more nuanced position of "deliberately unprofitable in this specific circumstance for this specific reason for this specific duration." This is the answer worth probing. The accounts run by agencies in this third category are leaving money in the business, sometimes a lot of it, by refusing to use one of the most powerful instruments available to them.

Deliberate-loss advertising is a discipline, not an excuse. It requires more rigour than standard performance management, not less. It requires more finance integration, not less. It requires more reporting structure, not less. The agencies that do it well are doing harder work than the agencies that don't. The operators that benefit from it are getting more from their ad spend than the operators who insist that every pound must be locally profitable.