The short answer: A vendor produces what you asked for. An advisor decides what should be asked for, and most agreements quietly price only the first. Sort last quarter’s deliverables into production and decisions. In one B2B archive we audited, 708 published articles had accumulated while the highest-traffic pages all sat off the lane the business actually sells in.
This is written for the person who signs the marketing partner’s invoice, whatever industry sits behind that signature
.
Every fall the same meeting happens. Someone pulls up a spreadsheet of what the agency delivered, counts it, and tries to work out whether it was worth the money. Volume is the only thing on the page, because volume is the only thing that was ever counted. A CMO Barometer survey of 805 marketing leaders, published in November 2025, found 69% rank creativity as the top capability they expect from an agency in 2026. Only 54% rank proactivity, the trait that separates a partner who tells you what to do next from one who waits, as high a priority.
Allied Insight runs its own client archive through the same audit it recommends to readers, and the pattern holds on our own site too. Publishing volume did not buy on-lane position. After 708 pieces, the top-traffic keywords on our own resource hub are still generic SEO and social-media terms, not the lane the business actually sells in. A thin on-lane head sitting under a large off-lane tail, measured on our own archive, not a client’s.
This guide covers how to sort your own last quarter into produced and decided, four signals that the decision half of the job has quietly stopped, and what to track for one quarter. It draws on patterns across our client portfolio.
What is the difference between a marketing vendor and a marketing advisor?
A vendor executes a brief someone else wrote. An advisor writes the brief, and most contracts only bill for the execution. That is the whole difference, and it has nothing to do with talent or price. A skilled team can build strong work from a bad brief all year and never once ask whether the brief was right.
An advisor’s actual job is asking that question first, out loud, before the work starts. The confusion happens because both jobs use the same vocabulary. A vendor and an advisor both say “strategy” in the kickoff call. The vendor means the plan for executing what was already decided. The advisor means the process of deciding what to execute in the first place. Nobody defines the word going in, so nobody notices which one they bought until a year has passed and the business looks the same as it did in January.
The relationship math makes the confusion durable rather than self-correcting. According to the 2025 ANA and 4A’s Client-Agency Relationship Tenure Report, the average client-agency relationship now runs about seven years, more than double the 3.2-year average measured in 2016. A vendor relationship that produces reliably gives an owner no obvious reason to reopen the question. The work keeps arriving. Nobody asks who is deciding what work to make.
Why does a partner stop making decisions?
A partner stops making decisions the moment nobody checks whether they still are. It rarely happens on purpose. Production has a visible failure mode: a missed deadline, a typo, a broken link. Decision-making has none. A quiet quarter looks identical to a good one, right up until a competitor’s move makes the gap visible.
An AI answer engine rarely cites a homepage or a generic services page by name. It cites a narrow page built to settle one reader question in the opening minute, and most firms have written surprisingly few of those, because writing one means deciding which question earned that focus. Deciding is not the same job as publishing, and deciding is what quietly stops once a relationship settles into a rhythm.
The Gartner 2026 CMO Spend Survey polled 401 marketing leaders across North America, the UK and Europe. It found that 56% of CMOs say their organization lacks sufficient budget for 2026 strategy work. Separately, 70% say internal processes are not mature enough to scale the tools they already have. Those are internal numbers, not agency numbers, but they describe the same vacuum a partner can quietly fill or quietly abandon.
Read your own invoice: the production column and the decision column
Pull the last three months of invoices and sort every line into two columns: things produced, and decisions made. Most businesses have never done this, because an invoice justifies a cost, not who is thinking. The sort takes about twenty minutes and reads the relationship more accurately than any pitch deck.
Allied Insight’s own pipeline-discipline audit work starts from the same premise: volume on a report and progress toward a goal are two different measurements, and confusing them is what lets a relationship coast. Three things make the sort useful rather than just tidy.
First, look for the counterintuitive line item. A production column is full of expected things: posts, pages, ads, emails. A decision column has an item that would have surprised you a year ago: stop doing something, retarget a page, kill a campaign that was technically working but strategically wrong. If every line in the decision column is a version of “more of what we already do,” the column is padded, not real.
Second, name the doctrine behind at least one decision. An advisor who is actually deciding can tell you the rule they applied and why, not just the recommendation. “Budget moved away from that channel because engagement rose while conversions flattened” is a doctrine, because it names the split. “We recommend increasing spend there” is a production instruction wearing a strategy costume.
Third, check whether anyone closed the loop. A real decision gets revisited: did the change actually move the number it was supposed to move. A decision from two quarters ago with no follow-up note anywhere was made once and abandoned. That is close enough to never having been made.
Pull your last quarter into two columns: produced, and decided. A working partnership fills both. A thin decision column means you are buying output and calling it strategy. Not sure which column a line belongs in? Compare notes with us
Four signals that the strategy work is happening somewhere else
Four patterns show up reliably once the decision half of the job has moved to someone else, and none of them need a forensic audit to spot. This is a starting checklist, not a diagnosis. A firm can show one or two signals and still have a working relationship.
Three or four together are the pattern worth acting on.
The table below sorts what each signal looks like on the surface from what it usually means underneath. Read it against your own last quarter before you decide anything.
| Signal | What it looks like on the surface | What it usually means |
| Every recommendation defends existing spend | Reports emphasize what is working, rarely what should stop | Nobody is being paid to find the weak line item |
| Strategy conversations happen in the kickoff, never after | The annual plan gets built once, then executed on autopilot | The advisor role only activated for one meeting a year |
| You supply the direction, they supply the execution | Briefs originate from your team, not theirs | The billing says advisor, the workflow says vendor |
| Reporting counts activity, not outcomes | Decks lead with volume: posts published, emails sent, ads run | There is no answer ready for “so what changed because of this” |
One field note on why the second row gets missed longest: an annual planning session feels like strategy, a room, a deck, a plan for the year. But a plan built once in January and never revisited is a production schedule wearing a strategic label. The advisor’s real job is the eleven months after that meeting, not the meeting itself.
The offensive play most firms miss: ask your partner, unprompted, what they would stop doing if the budget shrank 20%. A vendor answers with what is easiest to cut. An advisor answers with what matters least. That is a different list, and the gap between the two is usually where the decision work went missing.
Ask your partner what they would cut first if the budget shrank 20% tomorrow. A working program answers without a week to prepare. Silence means nobody has run that diagnosis. Want a second read on your budget? Ask us what we’d check
What to track for one quarter before you decide anything
Track three numbers for one quarter before renewing, replacing, or renegotiating anything: decisions proposed, decisions actioned, and the gap between them. Three metrics, not ten. A business owner running this diagnostic alongside an actual job needs something they can hold in their head.
Decisions proposed is the count of genuine strategic recommendations your partner puts in writing in a quarter, the kind that would land in the decision column from the invoice exercise above. A healthy number is at least one a month. Zero in a full quarter is the clearest single signal in this guide.
Decisions actioned is how many of those actually got implemented, by either side. A large gap between proposed and actioned can mean two different things, and the diagnostic only works if you tell them apart. If decisions are proposed but not actioned, the problem usually sits on your side: nobody internally owns saying yes. If decisions are neither proposed nor actioned, the problem sits with the partner. The advisor function was never running in the first place.
Time to close the loop is how long it takes, after a decision is actioned, for someone to report back on whether it worked. Automate the timing of that check-in rather than relying on memory: a calendar reminder tied to the decision date, never a human recalling a conversation from ten weeks back. A partnership that cannot close its own loop in under a quarter is running on the production half alone, whatever the contract says.
Track these three starting now, before the conversation about renewing or replacing anyone happens. Firms that time their end-of-year planning around a season, not a calendar page, run the same discipline: check the number before the meeting, not during it. A firm making the case for a new partner with no baseline is negotiating from a feeling. A firm defending its current one with no baseline is doing the same thing from the other side of the table.
Frequently asked questions
The two jobs, and who is doing which one
A marketing partnership is two jobs wearing one invoice: the work that gets produced, and the decisions about what should get produced. Only one shows up automatically on a spreadsheet, which is why the other goes missing without anyone deciding to cut it.
The sort in this guide, twenty minutes against your own last quarter, tells you which job you have actually been buying. Allied Insight builds the decision half of that job deliberately, through monthly marketing programs structured so the question keeps getting asked instead of quietly dropping. If your own invoice sort came back thin on the decision side, see what that looks like for your business.
References
- “ANA & 4As Client-Agency Relationship Tenure Report.” Association of National Advertisers and the 4A’s. Published April 2025. Verified live 2026-09-16.
- “Gartner 2026 CMO Spend Survey Finds CMOs Allocate 15.3% of Marketing Budgets to AI, But Only 30% Are Ready to Scale AI Capabilities.” Gartner. Published May 11, 2026. Verified live 2026-09-16.
- “What Global CMOs Want in 2026: Strategic Priorities and Agency Expectations.” CMO Barometer, GlobeNewswire. Published November 26, 2025. Verified live 2026-09-16.
- Allied Insight client engagement data, on-lane versus off-lane content performance, 2026 (anonymized and aggregated).
