An account manager gets a quarterly target for account growth and pitches every client the same three add-ons from a price list. One buys, two go quiet, and the monthly review starts to feel like a sales call. The upsell that keeps clients works the other way round: it starts from a problem the client's own data already shows, so even a "no" leaves them better informed than before.
An upsell is a finding with a price
Every add-on worth offering has three parts: a trigger, an offer, and a proof. The trigger is evidence in data you already collect, the audit, the rank history, Search Console. The offer is a named package with fixed deliverables and a price, built the way Productising SEO services builds them, because "some extra local work" is not something a client can approve. The proof is the measurement you commit to showing afterwards.
If you can't point at the trigger, the pitch is a guess. If you can't name the proof, you're asking the client to buy on faith twice, once now and once at renewal. The natural venue is the monthly review, after the standing agenda: the client has just seen the data, so the finding needs no wind-up.
The local add-on
The trigger: the client serves a geographic area, the map pack shows on their money searches, and their visibility varies sharply by neighbourhood. Local results depend partly on where the searcher is standing, so the single position a tracker reports hides the real picture: strong two streets from the office, invisible across town where the customers they want actually live. The free Local rank grid makes that variation visible on a map.
The offer is a local package: Business Profile management, a review cadence, and location pages for the areas the grid shows as weak. The proof is the same grid, re-run after the work. If local clients are new ground for you, How local search works opens that playbook, and Local and e-commerce playbooks shows how to keep it on file.
The content add-on
The trigger sits in Search Console: queries where the site earns impressions but has no page that really answers them, or rankings clustered on a few service pages while the questions buyers ask before choosing get no coverage at all. That gap is demand the client is already being shown for and not capturing.
The offer is a content package on a calendar the client approves, planned the way Content calendars that ship lays out. For a store the gaps usually sit at collection level and in buying guides, which is a different playbook: Why some stores rank and yours doesn't opens it. The proof is the target queries joining your tracked list and climbing, and the new pages collecting clicks in Search Console.
AI visibility, technical work, and when to stay quiet
The AI visibility add-on often triggers itself: the client asks whether ChatGPT mentions them, or AI-generated answers start appearing on their queries. Offer measurement first and improvement second, reported the way Reporting AI visibility describes, measured mentions only, no invented scores. The field is young, so sell the measuring and the work, never a promised placement.
The technical add-on is usually a project rather than a retainer line: a migration with a redirect map, a rebuild, speed work, or fixing whatever the audit keeps flagging beyond the retainer's scope. Its proof is the cleanest of all, a re-audit passing checks that used to fail.
And sometimes the right move is no pitch. Skip the upsell when there's no trigger in the data, when the last review carried bad news and the recovery plan hasn't shown results yet, or when the add-on solves your revenue problem rather than the client's. The test that keeps you honest: would you still recommend this work if the client did it in-house, with nothing in it for you? If not, it isn't a recommendation, and clients can tell the difference.
What to take away
- An upsell that retains has a trigger in the client's own data, a named offer with fixed deliverables, and a proof you commit to before the work starts.
- The local trigger is visibility that varies across the map, the content trigger is demand in Search Console with no page to capture it, and the AI trigger is usually the client asking first.
- Offer add-ons inside the monthly review, right after the data that justifies them, and never in a bad month before the recovery plan has shown results.
- If you wouldn't recommend the work with no revenue attached, don't pitch it.
Next
More add-ons mean more delivery, so the next chapter is about doing bigger work with the team you have: start with AI in the agency workflow.