Planning 2026 paid budgets around five known changes and a reserve for the rest
Apple, Google, Meta and the courts have all moved this year. Here is how to build a 2026 budget that absorbs known changes and keeps room for ones we cannot see.
A 2026 budget built by rolling last year's split forward will be wrong in specific, predictable ways, because several things that shape cost and measurement have changed this year. The goal is not a forecast of 2026. It is a plan with a stable core, a funded set of tests, and a reserve with rules for using it. Here are the five known changes we would plan around, followed by the structure.
Five changes that affect the plan
- Apple Ads will show more ads in App Store search results. Reporting on Apple's announcement this week says ads will appear at the top of results and further down the page beginning in early 2026, supported on iOS and iPadOS 26.2 and later. Existing search campaigns are eligible without changes, billing stays cost-per-tap, and Apple's help page says you cannot select or bid for a particular ad position. The date has not been confirmed by Apple in what we read. Plan for a change in impressions, tap volume and cost per tap, and do not assume which way.
- Google's AI answers carry ads you cannot isolate. Ads can appear within AI Overviews, you cannot opt out, and Google Ads documentation says reporting for them is not segmented. Search results pages are changing underneath your Search budget, so treat shifts in click-through and conversion rate with caution.
- The browser privacy roadmap is settled, for now. Google said in October it is retiring most Privacy Sandbox technologies, including the Attribution Reporting API and Topics, after earlier deciding to keep Chrome's current approach to third-party cookie choice. There is no browser-level substitute coming for your own measurement.
- US app marketers have a new route, with an unfinished commercial term. An April court order allowed apps on the US storefront to link out to web purchases. On December 11, the Ninth Circuit reversed the total ban on Apple charging a commission for those purchases, but Apple cannot charge anything yet. A fair rate must be agreed with Epic or set by the court. Treat web checkout economics as unsettled until that is resolved.
- Meta's automation direction is stated. In June the Wall Street Journal reported that Meta aims to let brands create and target ads fully with AI by the end of 2026. That is a report of an aim. Budget for continued automation of delivery, and keep your conversion signals and independent measurement in good order.
A three-part structure
Core, about 70 to 80 percent
Fund what is working at levels your payback target supports, at the current mix. Do not rebalance the core on the basis of announcements. Rebalance on measured marginal returns. If you do not have them, the first investment of the year is a measurement project.
Tests, about 10 to 20 percent
Fund named experiments with a question, a budget, a duration and a decision rule. Candidates drawn from the changes above: new Apple search placements once they appear, Google's AI-driven Search features against a control, a US web-checkout flow with a holdout, and a creative-concept test on Meta. Write down in advance what result would make you scale and what result would make you stop.
Reserve, about 5 to 10 percent
Hold unallocated funds with triggers for release, such as a channel's cost per result moving beyond a set band for two weeks, or a test hitting its threshold. A reserve with no rules gets spent on whatever is loudest in January.
A hypothetical on a $1,200,000 annual budget: $900,000 core, $200,000 tests, $100,000 reserve. The percentages are a starting point, not a standard. A business with tight cash and a short payback window should shift toward the core.
Schedule reviews around the dates you can predict
- A review in the first weeks after Apple's new placements go live, comparing impressions, taps, cost per tap and downstream conversion against your last quarter.
- A review when the commission question for web purchases is resolved, if you are in the US test.
- A quarterly check on incrementality results, before budget moves between channels.
Write down what you would need to believe
For every channel, record the one or two assumptions that justify its share, such as "paid social returns at least 1.5 over 60 days" or "branded search is not simply capturing demand created elsewhere." Then each quarter ask whether the evidence still supports them. This keeps the plan honest when attribution is partial, and it makes reallocation a decision about evidence, not about enthusiasm.
The practical takeaway: split the budget into core, tests and reserve before the year begins, and attach a written decision rule to every dollar that is not core.