Hire the Retail Media Strategist Who Audits What the Bidding Agents Did

What Does a Retail Media Strategist Do Once Agents Place the Bids?

On Tuesday the dashboard says sponsored products beat their return target. On Thursday your largest advertiser calls to ask why their own brand keywords suddenly cost half again what they did last month. Both things are true, and the agent that caused the second was optimizing correctly for the first. Somebody has to own that gap, and that person is the retail media strategist.

Retailers have spent the last few years turning first-party shopper data into advertising platforms, and the named ones are now large businesses in their own right: Walmart Connect, Target's Roundel, Amazon Ads ¹. The execution layer of that business is going to software. Agents test promotions, adjust retail media bids in real time, and manage product feeds to improve discoverability across channels ¹. None of that is the part you are hiring for anymore.

What is left for a person is narrower and heavier. Pricing and packaging the network, so a category with thin inventory does not get sold as if it were endless. Deciding which advertisers get which placements when two of them want the same shopper at the same second. Writing the rules the agents run inside, and then checking whether the agents stayed inside them. The strategist is spending other companies' money at machine speed, which makes the audit trail the job rather than the paperwork around it.

The traits that matter are visible in how a candidate talks about a bad week. Real ones narrate a specific outage: the day of the month, the campaign type, the metric that looked fine while the problem grew underneath it, the advertiser they called before the advertiser called them. Performed ones narrate a philosophy of automation. Ask a follow-up about the mechanism, one level down. Someone who genuinely ran this can tell you which lever they pulled first, and why they did not pull the obvious one.

Which Backgrounds Produce Somebody Who Calls the Advertiser First?

The people who make that Thursday call before the advertiser does tend to arrive from a short list of seats. Marketplace and search ad operations, where somebody already ran sponsored placements against a catalog. Category management and trade promotion planning inside a consumer brand, where somebody already negotiated for shelf and funding. And retailer-side merchandising or pricing analytics, where somebody already knows why the margin on a promoted item is not what the promotion sheet says.

The first group brings speed and platform fluency, and often arrives short on commercial nerve: they can defend a bid strategy and cannot defend a rate card. The second group brings the opposite, and their advantage is underrated. A trade promotion planner has spent years arguing about who pays for a discount and what it did to baseline sales, which is exactly the argument a retail media network has with its advertisers, only faster. The third group brings suspicion of clean numbers, which is the trait that catches an agent quietly harvesting sales that were going to happen anyway.

The unexpected backgrounds are worth a call. Revenue management from hotels or airlines transfers almost directly: perishable inventory, dynamic pricing, automated systems making thousands of small allocation decisions, and a human accountable for the ones that embarrass the company. Product feed and catalog operations, the people who spent years fixing broken attributes so items would show up at all, understand the input layer the agents now manage. Internal audit and finance analysts sometimes make excellent hires here for one reason: they are comfortable asking a system to prove a number, and they do not stop asking when the number is good.

Much of this overlaps with what an AI-fluent marketing manager needs, and with the automation instincts of a GTM engineer. The distinguishing requirement is fiduciary. This role spends money belonging to advertisers who can leave.

What Did the Machine Get Confidently Wrong, and When?

Not by attending the training. The strategists who are genuinely good at this built a private habit of checking the machine, usually because it burned them once. Ask when that happened. The answer is often a specific number that looked plausible, was quoted to somebody senior, and took two days to unpick: the kind of number that reads fine on Tuesday and produces a phone call on Thursday.

What the practice looks like in this job is unglamorous. Before running an agent against a live budget, they ran it against a period they already understood, and compared what it would have done against what actually happened. They keep a small set of checks that do not come from the platform: a supplier's own sell-through, a store-level number, a query report read by hand on Monday morning. They know which of the platform's metrics are measured and which are modeled, and they say so unprompted when quoting one.

They are also specific about what they do not delegate. Rate cards, exclusivity, incrementality claims made to an advertiser in writing, and any decision that changes what a shopper sees for reasons the shopper would object to. That last one is judgment rather than policy, and candidates who have never thought about it will say the agent handles it.

Fluency without friction is the thing to distrust. If a candidate describes months of daily AI use and has no story about an output that was confidently wrong, either the tools have not been near anything consequential or the checking is not happening. Skepticism about a specific model behaviour, described in ordinary language, is worth more than a list of tools. This is the same evidence problem you hit hiring an AI red team engineer: the resume shows exposure and never shows practice.

Where Do Retail Media Strategists Who Already Run Agents Spend Their Time?

Roughly in order of yield, they sit inside the platforms themselves: Walmart Connect, Roundel, Instacart Ads, Amazon Ads and their retailer equivalents ¹. Inside the tooling vendors whose products sit between advertisers and those platforms, which is where Pacvue, Skai, Criteo and Stackline recruit and lose people. Inside agency commerce practices. And in the operator seats at brands large enough to run their own retail media spend.

The vendor and agency populations are the most reachable, because the work is portable and the people are visibly named on case studies and conference panels. Events worth working rather than attending: Shoptalk, eTail, Amazon Ads unBoxed, and the Path to Purchase Institute programming, where the retail media track is now the busy one. The IAB's retail media standards work draws in the practitioners who care about measurement definitions, which is a useful filter for the audit-minded profile described above.

Adjacent titles that convert well include retail media manager, marketplace or commerce lead, category manager on the brand side, and revenue or yield manager from a different industry. Search postings for the work rather than the title. A job that mentions bid automation, feed management and advertiser reporting in the same paragraph is this job, whatever it is called.

One warning about sourcing by pedigree. The largest networks have deep bench strength on execution and thin bench strength on network economics, because the economics were set above them. A senior title at a famous network does not tell you the person ever priced anything. Ask what they set, not what they ran.

How Do You Screen and Close a Retail Media Strategist Without Guessing?

Screen with a real artifact and a real constraint. Hand the candidate an anonymized week in which the return target was met and one advertiser's brand keyword costs climbed by half, give them an AI assistant, and give them forty minutes. What you are watching for is whether they frame before they generate, whether they ask the assistant for the source of the number that matters, and whether they catch the claim that is confidently wrong.

The usual interview substitute is a case study answered aloud, and it rewards articulacy about method rather than method. A candidate can describe an incrementality check beautifully and never have run one. Watch instead for the small refusals: declining to quote a modeled metric as measured, flagging that the sample week is too short for the conclusion, saying plainly that they would call the advertiser before changing the bid caps.

On closing, this population cares about authority and about measurement, and says so late. Authority means the rules the agents run inside: a strategist who discovers after starting that pricing and guardrails are owned by a different team will leave inside a year, and they will tell that story at conferences. Measurement means the network's own numbers. The offer dies quietly when a candidate concludes during the process that the reported returns would not survive scrutiny, so let them ask the measurement questions and answer them directly.

Smaller things kill offers too. Being buried under a media sales organization with no product voice, which the strong candidates read as a demotion. And an unclear line between this role and the seller who carries the advertiser relationship, which is a boundary worth writing down before the offer rather than after. If your network sells through named reps, decide who owns the account conversation and say it out loud, the same way you would when hiring an account executive.

A Bonus on Network Revenue Buys You a Worse Audit

No published salary series exists for this title yet, so treat any precise band you are quoted as somebody's sample rather than a market. What can be sourced is the adjacent base. Built In's 2026 US marketing manager page reports an average base of about $94,800, a median of $89,000, and a reported range from $36,000 to $230,000 ³. That range is wide because the title covers a two-person team and a platform business alike.

The premium for AI-capable work is measurable at the market level even where role-specific data is missing. Lightcast reported in July 2025 that postings including AI skills advertise about 28 percent higher salaries, close to $18,000 more per year ². As of mid-2026 the honest framing for a retail media strategist is the marketing manager band, shifted up by the network's revenue scale and the size of the spend under the person's control, with the AI premium sitting inside that shift rather than on top of it. That last part is a stated bet rather than a sourced number, and you should test it against two or three live offers before you build a range on it.

Comparable in-house roles at large networks frequently carry a bonus tied to network revenue. Decide early whether this role gets one, because a strategist paid on network revenue and asked to police the bidding agents has been handed two jobs that pull against each other. If you want the audit to be real, do not pay only for the number the audit is checking. The strategist whose bonus rides on network revenue is exactly the one who will not make the Thursday call.

On location, the work is hybrid in practice. Feed operations, reporting and agent supervision are genuinely remote-capable and often run distributed. Network pricing, advertiser escalations and the merchandising conversations that decide what may be promoted still cluster where the retailer's commercial teams sit, which is usually a headquarters with three days on site. Fully remote is workable for a strategist joining a vendor or agency and difficult for one joining a retailer's own network in its first two years, when almost every decision requires somebody from merchandising in the room.

See how it works >