4 Analytics Checks to Run Before You Start Planning Your Fall Campaigns

Open your August numbers and pull two figures. Sessions that added something to the cart. Sessions that completed a purchase. Subtract the second from the first, then divide by total sessions. For most stores the answer lands somewhere between 3% and 6%. Call it your Discountable Share. That is the entire slice of your traffic a fall discount can physically convert, and most fall promo calendars get priced as though it were the whole store.

Fall planning almost always runs in one direction. Someone opens a calendar and blocks out Labor Day, a mid-September push, a pre-Halloween moment, a Q4 warm-up. Then the analytics get opened, but only to support what is already drawn. It feels rigorous because numbers were involved. The problem is structural. A planning session with no veto can only grow. A fall planning session that ends with more promotions than it started with did not use data, it used enthusiasm. The only honest output of a pre-planning analytics check is a campaign you decide not to run.

The four checks below are built to say no. Each takes a few minutes, uses a report you already have, and ends in a cancel rule instead of a number to watch. By the end you will know which of your sketched fall campaigns should not survive the week.

Check One: Most of Your Traffic Is Not a Discount Problem

A discount acts on exactly one part of the funnel. The gap between wanting the item and completing the purchase. Every leak before that point is untouched by whatever percentage you pick. Your blended conversion rate hides which stage is leaking, which is why merchants plan promotions against a number no individual visitor ever experiences.

Run the Discountable Share on your own August

Take a store with 40,000 August sessions. Illustrative, but the shape is common. 3,600 of those sessions added something to the cart. 1,400 completed a purchase. The Discountable Share is 2,200 divided by 40,000, which is 5.5%. The other side of that number: 36,400 sessions never reached a cart at all. No markdown reaches them, because they left before price was ever the question. This is why a sitewide fall banner so often produces a revenue bump that evaporates on the margin line. It hands a discount to 100% of buyers in order to act on 5.5% of visitors, and part of that 5.5% was coming back anyway.

Read the funnel stage by stage, not as one number

Separate five stages: session start, product view, add to cart, checkout begin, completed order. Each drop has a different cause, and only one of them answers to price. Losing people between session start and product view is a navigation problem. Losing them between product view and add to cart is a belief problem: the photos, the sizing details, the reviews, or the value framing are not doing enough. Neither gets fixed by 15% off. Run the promotion anyway and you are paying to move a number you were never touching.

Cancel rule: if your largest stage-to-stage drop happens before add to cart, the fall discount campaign comes off the calendar and the budget moves to the stage that is actually leaking.

A discount is a checkout-stage instrument. Your Discountable Share is the ceiling on what any fall promotion can deliver, and in most stores that ceiling is about 5%. A sitewide markdown is paid by 100% of your buyers. Every fall calendar built on one percentage treats those two numbers as if they were the same one.

Check Two: Your Sessions-to-Purchase Report Already Wrote Your Campaign Length

Campaign duration usually gets picked for tidiness. Twenty-four hours. A weekend. A week. Almost nobody derives it from how long their buyers actually take to decide. Two numbers already answer that question: sessions to purchase and days to purchase. Read them as a distribution, not as an average.

The 48-hour flash sale problem

Say your report shows a median of 4 sessions across 9 days for a $160 home goods item. A 48-hour flash asks a nine-day decision to compress into two. Almost nobody does that. What happens instead is that shoppers already on session three convert, get counted as campaign performance, and make the flash look like a win. Meanwhile everyone on session one sees a deadline they cannot meet, files your store under "runs sales," and waits for the next one. The campaign did not create demand. It harvested the shoppers who were nearly there and taught the rest to wait.

Use the median, then look for the second hump

Averages lie badly here. A handful of shoppers who took 40 days drags the mean upward and makes your buyers look slower than they are. Use the median. Then look for a second cluster. Plenty of stores have two populations hiding inside one report. Repeat buyers who convert in one or two sessions within a day. Considered purchases that take a week or more. One campaign length serves one group and quietly misses the other. Once you can see the split, the question stops being "how long should the sale run" and becomes "which of these two groups is this campaign for."

Most merchants skip these first two checks because building them by hand takes an afternoon. Growth Suite reports the funnel stage by stage, from session start through completed order, and reports purchase insight directly: how many sessions and how much time pass between a shopper's first visit and their order, plus cart-level detail on what stalls. That turns Check One and Check Two into a five-minute read instead of a spreadsheet project.

Cancel rule: if your median buyer needs more sessions or more days than your planned window allows, rebuild the window or cut the campaign. Campaign length is not a marketing preference. It is a number already sitting in your report, and picking a round one instead is how a fall sale ends up talking to the shoppers who needed it least.

Check Three: The Products You Want to Feature Have the Least to Gain

Fall campaign briefs default to best sellers. Best sellers sit at the top of the revenue report, so they are what everyone sees first. They also already convert. Discounting them buys volume you had and surrenders margin you were keeping. That is the single most expensive default in seasonal planning.

The overlap test takes two minutes

Export the product-level report and sort by revenue. Put the fall campaign brief next to it. Count the overlap. In most planning documents three or four featured products are also top five by revenue, which means the campaign is built to cut the price of your most reliable full-price demand. Run the math once. Your top seller does 300 units a month at a 52% margin. Cut 20% off for a two-week push and profit per unit falls from 52 cents on the dollar to 32. To hold the same gross profit you now need 488 units instead of 300. That is a 63% unit lift on a product that was selling fine. Revenue will still go up. The profit line is where the bill shows up.

Traffic without conversion is where a campaign belongs

Pair each product with two numbers instead of one. How many people viewed it, and what share of those viewers bought. High views with low conversion is the most useful pattern in the report. Shoppers are finding the product and deciding against it. That is a fixable objection, and the one place a promotion creates a sale that would not otherwise exist. Low views with high conversion is a different problem. Those products do not need a discount, they need to be seen. Putting them in a sale treats a merchandising gap as a pricing gap.

Doing this pairing by hand across a few hundred SKUs is why the check gets skipped. Growth Suite segments products on exactly this logic, sorting the catalog into buckets like Stars, Stoppers, Invisibles, Gems, and Underperformers based on how traffic and conversion interact, with a product-level report and CSV export behind it. The fall feature list then gets picked from products where a promotion can change an outcome, and your Stars stay at full price where they belong.

Cancel rule: line up your top five revenue products against your fall feature list. If more than one appears on both, rewrite the brief before anyone writes the creative.

Your best sellers have the least to gain from a markdown and the most margin to lose. Featuring them is the default move in fall planning, and the default is the expensive one.

Check Four: When Abandoned Carts Are Bigger Than Orders, Price Is Not the Blocker

Two numbers you almost never see side by side: the average value of your abandoned carts and the average value of your completed orders. Pull both. When abandoned carts run consistently larger, the shopper's problem is the total, not the item price, and a percentage off is the wrong instrument.

What the gap actually tells you

Suppose completed orders average $78 and abandoned carts average $124. That gap is a message. Shoppers are building a basket they want, then reacting to what the total looks like at the moment of commitment. Usually right after shipping and tax appear. A 15% discount takes that $124 cart to $105, still well above the number they were comfortable with. So the promotion pays margin on every order without moving the ones it was designed for. A free shipping threshold set just under $124 does something different. The basket they already built clears the bar, the shipping line disappears, and the total drops without a markdown on a single product.

Choose the instrument before you choose the number

The order of operations matters here. Most fall planning picks a percentage first and then hunts for somewhere to apply it. This check reverses that.

1. Read the objection: does the cart data say shoppers are reacting to the item price or to the total?
2. Pick the mechanic: percentage off, a spend threshold, a bundle price, a volume tier, or clearer delivery timing at the cart.
3. Then set the number: and not one minute before.

Sometimes the answer is still a discount. Often it is a threshold, a bundle, or simply showing delivery dates at the cart stage, which costs no margin at all. Cancel rule: if the abandoned cart gap is wide, replace the planned percentage-off campaign with a threshold mechanic and hold it to the same target.

When abandoned carts are consistently larger than completed orders, the shopper is reacting to the total, not the price. A percentage off treats a threshold problem as a desire problem and pays for the mistake in margin.

What Percent of Your Traffic Can This Discount Actually Reach?

Four checks, four cancel rules. Your Discountable Share is single digit, so a fall promotion aimed at everyone is priced against a population it cannot reach. Campaign length belongs to your sessions-to-purchase distribution, not to a round number on a calendar. The products at the top of your revenue report have the least to gain from a markdown, and featuring them is the most common way to end a quarter with more orders and less profit. And a wide gap between abandoned cart value and order value means the objection is the total, which a threshold fixes and a percentage does not.

Before you write a single line of your fall calendar, run the first check. Pull August sessions, cart-adds, and orders, and calculate your Discountable Share. It takes two minutes and it reprices the whole plan. If the number surprises you, every other assumption on the whiteboard deserves the same scrutiny.

If your fall calendar is a stack of campaigns nobody has tested against the numbers, Growth Suite helps you tell walk-away customers apart from dedicated buyers and show one genuine, time-limited offer only to the visitors likely to leave without purchasing. So you convert your Discountable Share without paying the shoppers who were already going to buy. It is free to install on the Shopify App Store, with a 14-day free trial.

Frequently Asked Questions