Use Meridian Scenario Planner to Defend Your 2026 Shopify Budget Shift

Budget Fights Need Better Inputs

The first 2026 budget fight I would prepare for is the boring one: Google Ads wants another 18%, Meta wants its December spend protected, SEO wants patience, and the founder wants to know why blended CAC moved from $42 to $51 after Black Friday. The room does not need another channel screenshot. It needs one model, fed with revenue the finance team trusts and conversion signals the marketing team can explain.

That is where Google’s Meridian Scenario Planner gets useful. Google announced Scenario Planner on February 19, 2026 as a marketer-friendly interface for testing budget scenarios from a trained Meridian marketing mix model. The official docs call it an open beta tool that creates reports with model fit, channel contribution, return on investment, and interactive budget optimization from a model stored in Google Drive or Google Cloud Storage. You can read Google’s announcement here: New Meridian tool puts MMM insights directly in marketers’ hands. The docs are here: Meridian Scenario Planner.

The planner is no-code at the decision layer. That does not make the work no-ops. For a Shopify brand doing $4 million to $40 million a year, the win comes from cleaning the Shopify revenue export, separating promo pressure from media pressure, and tying GA4 key events to the purchase path before anyone starts dragging budget sliders around.

Start With Shopify Revenue, Not Platform Revenue

I do not use ad platform purchase value as the outcome column for an MMM. Google Ads, Meta Ads, TikTok Ads, and Klaviyo all tell useful stories, but each one has its own attribution window and self-interest baked into the dashboard. Shopify is closer to the cash register. It has refunds, discounts, taxes, shipping, and order-level timing in the same system that operations already reconciles.

Shopify’s order export gives you order fields such as Name, Financial Status, Paid at, Currency, Subtotal, Shipping, Taxes, Total, Discount Code, Discount Amount, and Created at. Shopify also notes that exports over 51 orders or date-based exports are emailed to the store owner, and that very large exports can take hours. Their example says fewer than 100,000 items may finish in under an hour, while 400,000 items may take around 4 hours. Source: Shopify order exports.

For Meridian, I would build the weekly outcome from Shopify finance or order data, not from raw line items pasted into Sheets at 11 p.m. Pick one revenue definition and write it down. Shopify’s finance docs define net sales as gross sales minus discounts and sales reversals, excluding shipping and taxes. Total sales adds taxes, shipping charges, and fees. Source: Shopify finance reports. For budget planning, I prefer net sales before tax and shipping for ecommerce MMM because media does not create sales tax, and shipping revenue can move when the ops team changes a free-shipping threshold.

A small example makes the point. Say a skincare store booked $186,400 in Shopify total sales during the week of March 9, 2026. That number includes $11,200 of shipping charges and $8,900 of tax. Net sales were $166,300 after $24,600 in discounts and $6,800 in returns. If Meta claims $94,000 of purchase value for that same week, the platform number might still be useful for campaign QA, but it is not the business outcome I would feed into Meridian.

Promo Pressure Needs Its Own Column

Promotions are where channel politics get ugly. Paid search gets blamed when contribution margin drops, then someone remembers the 25% off code that ran from Thursday through Sunday. Meridian’s 2025 updates added support for non-media variables like pricing and promotions, which is exactly the opening Shopify operators need. Google’s September 30, 2025 update says Meridian can include pricing and promotions to measure their impact on sales, plus channel-level contribution priors and marginal ROI priors. Source: Meridian budget decision updates.

I would give promos at least three fields in the modeling table: discount dollars, promo flag, and promo type. Discount dollars can come from Shopify’s Discount Amount or the finance report’s discounts field. Promo flag is a clean 0 or 1 for each week. Promo type is plain language: sitewide_20, bfcm, free_shipping_75, bundle_3_for_2, or vip_early_access. If the store ran Klaviyo SMS early access on Tuesday and opened the same code to everyone on Friday, that belongs in the notes table too.

This matters in Scenario Planner because the budget slider is only as honest as the model behind it. If Cyber Monday revenue is treated as media impact when the actual driver was BFCM30, the planner may tell the team to pour 2026 dollars into the channel that happened to be loudest during the promo. I have seen this happen with branded search. The model sees a spike, the search team gets credit, and nobody notices that 42% of that week’s Shopify orders used the same discount code.

A practical rule: build one row per week, with Monday as the week start, and keep promo variables separate from spend variables. For a U.S. Shopify store, I would use the store’s reporting timezone, not UTC, because Shopify order timing, Klaviyo sends, and GA4 event dates drift if one system closes the day five hours before the others. The week of November 24, 2025 needs to mean the same seven days in every export.

GA4 Key Events Are Sanity Checks, Not Revenue Truth

GA4 changed the language around conversions. Google now calls important behavioral actions in Analytics key events, while Google Ads conversions stay tied to ads reporting. Google says existing Analytics conversions became key events and can still be created and reported in the same way. Source: Conversions vs. key events in Google Analytics.

For a Shopify growth team, I would pull weekly counts for a tight set of GA4 key events: purchase, begin_checkout, add_to_cart, generate_lead if wholesale or quiz capture matters, and maybe sign_up if email acquisition is a real program. Do not dump 27 events into the planning conversation because GA4 has them. If nobody would change 2026 spend because view_item_list moved 7%, it does not belong in the deck.

The useful move is triangulation. Shopify net sales is the outcome. GA4 purchase key events are the tracking health check. If Shopify says orders rose 14% in the week of May 4, 2026 and GA4 purchase key events fell 9%, you probably have a tagging or consent-mode shift to investigate before anyone trusts the model. If Shopify orders and GA4 purchases move together within 3% to 8% for most weeks, the data is good enough for planning. It does not have to be perfect. Perfect usually means someone rounded off the inconvenient bits.

I also like using begin_checkout as an early signal for organic and retention programs. Paid media can push sessions fast, but SEO content, email flows, and product launches show up in checkout starts before finance feels the full revenue effect. A founder may not fund organic from a last-click report, but Scenario Planner can make the case if the model sees the relationship between weekly organic sessions, checkout starts, and net sales over 104 weeks.

Build The Table Before You Open The Planner

The working dataset does not need to be fancy. A clean Google Sheet, BigQuery table, or CSV can carry the first pass. The row grain should be weekly. The columns should include week_start, net_sales, orders, discount_amount, promo_flag, google_search_spend, google_pmax_spend, meta_spend, tiktok_spend, affiliate_commission, email_sends, organic_sessions, purchase_key_events, and begin_checkout_key_events.

Keep media inputs boring and traceable. Google Ads cost should come from Google Ads or the MMM Data Platform if the account has access. Meta, TikTok, Pinterest, Reddit, Snap, affiliate, podcast, and direct mail spend should come from the platform invoice or the finance export, then mapped to the same week. Google’s 2026 Analytics announcement says enhanced budgeting tools in GA360 are meant to bring paid, third-party media like Pinterest, Reddit, Snap, TikTok, and organic performance into one measurement experience. Source: Enhanced budgeting tools in Google Analytics powered by Meridian.

Do two ugly checks before modeling. First, sum 2025 weekly Shopify net sales and compare it with the finance report total for January 1 through December 31, 2025. I want the variance under 1%. Second, sum each channel’s weekly spend and compare it with invoices. If Meta spend says $612,000 in the model table and finance paid $641,000, find the missing $29,000 before Scenario Planner gives you a precise-looking answer.

The model fit section in Scenario Planner shows expected versus actual outcomes plus fit metrics such as R-squared, MAPE, and wMAPE, according to Google’s docs. Those charts are useful, but I would not let a pretty fit metric end the conversation. A model can hug historical revenue and still be weak for causal planning if promo, inventory outages, or PR spikes are missing. Add a calendar column for stockouts, product drops, creator campaigns, and offline events. The planner cannot guess that the navy carry-on sold out for 18 days in July.

Use Scenarios To Make One Decision At A Time

The worst way to use Scenario Planner is to ask, “What should our 2026 budget be?” That question is too broad. Ask the version finance can approve on a Tuesday: “If we hold total spend at $1.2 million for Q2 2026, should $120,000 move from Meta prospecting into Google Shopping and YouTube?” Now the planner has a job.

I would run three scenarios for that question. The base case freezes 2025 channel mix and applies the 2026 budget. The operator case shifts 10% of spend toward the channels with stronger marginal ROI. The constraint case protects brand search, email, and affiliate while testing only prospecting dollars. In Meridian language, the room should care about marginal ROI because it answers where the next dollar has room to work, not which channel had the best average ROI after years of underfunding or overfunding.

Use real stakes. If the planner estimates that moving $80,000 from Meta to Performance Max raises Q2 revenue by $146,000 with a credible interval that crosses close to zero, I would not present it as a sure win. I would present it as a test budget with a stop-loss. Spend $25,000 in April, watch Shopify net sales, GA4 begin_checkout, branded search lift, and contribution margin, then decide whether the remaining $55,000 moves in May.

The same logic helps organic. SEO and lifecycle teams hate MMM rooms because spend is lumpy and effort is not always booked as media. Fine. Put the labor or agency retainer in the table. A $14,000 monthly content program, a $9,500 Klaviyo agency fee, and 38,000 non-brand organic sessions are concrete enough to model. Scenario Planner will not solve every attribution argument, but it gives the founder a way to compare paid media dollars with operating spend that used to hide below the line.

The Budget Memo I Would Bring To The Founder

My final memo would fit on two pages. Page one says what changed: 2025 Shopify net sales, total marketing spend, blended MER, discount rate, channel spend, and model-read incremental revenue. Page two says what I want to do in 2026: move a named amount, from named channels, during named weeks, with a measurement plan attached.

For example: “Shift $180,000 in Q2 2026 from Meta broad prospecting into Google Shopping, YouTube remarketing, and creator whitelisting. Keep total Q2 spend at $1.35 million. Hold sitewide discounts under 11% of gross sales except Memorial Day week. Review after the first four weekly reads, using Shopify net sales as the outcome and GA4 purchase key events as the tracking QA check.” That is a budget decision. Nobody has to pretend last-click ROAS is the judge.

Meridian Scenario Planner changes the meeting because it lets non-technical operators touch the tradeoffs themselves. The real work happens before the meeting, in the export logic, revenue definitions, promo calendar, and GA4 key event cleanup. Once those inputs are tight, budget debates get less personal. The question stops being which channel team has the loudest dashboard and becomes where the next dollar is likely to earn its keep in 2026.