Bundles and Good-Better-Best Prices: A Worksheet for Digital Product Sellers
About Alex: a pen name. There is no shop behind it, and no results are claimed. What follows is arithmetic and opinion, and I'll mark which is which.
In short
A bundle is a discount with a bow on it. Whether it is a good idea depends on how much profit one bundle sale leaves and on how many bundle buyers would have bought several of your products anyway. You can calculate the first exactly and only guess the second, so the guess goes into a written rule and then into a small test. (That framing is my opinion.)
What you get from reading it:
- A Bundle Worksheet: profit per bundle against separate sales, three price lines, and a try-or-not rule
- A good / better / best layout with what each tier must contain, plus a Tier Worksheet
- Two imaginary sellers, one where the bundle works on paper and one where it does not
You have three products, or five, or twelve. Someone says "put them in a bundle" and someone else says "offer three tiers". Both are easy to build in an evening. What is harder to see is the cost of getting them wrong: a bundle priced too deep can leave you less money for more work.
This guide does not say bundles sell more; I have no source for that. It shows what a bundle or a set of tiers leaves you per sale, so you can decide whether a test is worth an evening.
It builds on Price Your Digital Product So It Actually Leaves Profit. That guide explains profit per sale, margin, and the stand-still check for price cuts. I reuse its formulas here and won't re-teach the basics, so read it first if "profit per sale" is a new phrase.
Who this is for. A seller of digital products (templates, printables, planners, fonts) with at least three products on a marketplace or their own site. With fewer than three, a bundle has little to bundle.
What it costs. Nothing and no software; a notebook is enough.
A note on fee numbers. I quote no platform's fees. Every fee below is invented. Copy yours from your own fee page on the day you work, and write the date down.
What a bundle changes in the arithmetic
A bundle changes three things, and the worksheet follows all three.
- The price. You charge less than the items would cost separately. The gap is the discount.
- The fees. A percentage fee shrinks with the price. A flat fee may be charged once per order, not once per item, and the bundle then saves some of it. I don't know whether your platform works this way, so the worksheet treats it as an assumption to check against your fee page and a real payout.
- The buyers. Some bundle buyers would have bought only one item. Some would have bought all of them anyway. The first group is where a bundle can add profit. The second is where it takes profit away.
The sums can't settle the third point. I call it the would-have-anyway share and treat it as a guess to test.
Two imaginary sellers carry the examples. Kai sells three products. Mo sells three cheaper ones. Neither exists, and every figure attached to them is invented to show the arithmetic. Fees are invented too and are not any platform's.
The Bundle Worksheet
Copy the inputs from your records, then calculate A to G with the formulas below the box.
BUNDLE WORKSHEET date: ____
Fee page read on: ____
Product prices (singles):
P1 $ ____ P2 $ ____ P3 $ ____
(add more lines for more items)
1. Percent fees (all added): ___ %
2. Fixed fee per order: $ ____
3. Other cost per order: $ ____
4. Planned bundle price: $ ____
5. Singles you want to
replace (N): ____
A. Profit of each single
P1 $ ____ P2 $ ____ P3 $ ____
B. Average single profit (Ps)
= $ ____
C. All bought separately (T)
= $ ____
D. Bundle profit (Pb) = $ ____
E. Discount = ____ %
F. Bundle sales that replace N
singles, round UP = ____
G. Price lines 1, 2, 3 and c*
(formulas below)
The formulas, written out
single profit =
price * (1 - pct/100)
- fixed - other
Ps = (sum of single profits) / items
T = sum of single profits
Pb = bundle * (1 - pct/100)
- fixed - other
discount = 1 - bundle / sum of prices
If you are not sure your platform charges the flat fee once per order, run the worksheet twice, the second time with the flat fee and other cost multiplied by the number of items. If the answer you want survives only the first run, treat it as unproven.
Three price lines
The discount you can afford depends on what you compare the bundle with. I would use three lines, from loosest to strictest.
LINE 1 profit floor
bundle >= (fixed + other)
/ (1 - pct/100)
Below it, every bundle sale loses.
LINE 2 single-buyer line
bundle >= (Ps + fixed + other)
/ (1 - pct/100)
This equals the average single
price. Below it, a bundle sale
leaves less than one average
single sale.
LINE 3 parity line
bundle >= (T + fixed + other)
/ (1 - pct/100)
At or above it, a bundle sale
leaves at least as much as
buying all items separately.
Line 1 is the only line that says "still leaves a profit" and nothing more. A bundle can clear it by a mile and still be a poor trade. Line 2 protects you from a buyer who would have bought one item. Line 3 protects you from a buyer who would have bought them all. Between lines 2 and 3 sits the question of who your bundle buyers are.
The would-have-anyway share
Let c be the share of bundle buyers who would have bought all the items separately. Suppose the rest would have bought one item each. That is a simplifying assumption; real buyers vary. The profit change per bundle sale is:
change = (1 - c) * (Pb - Ps)
+ c * (Pb - T)
Set that to zero and solve:
c* = (Pb - Ps) / (T - Ps)
c* is the break-even share. If your honest guess for c is below c, the bundle adds profit under these assumptions; above it, it loses. This needs Pb larger than Ps; otherwise even the one-item buyer is worth less. If c is 1 or more, the bundle leaves at least as much as separate sales whatever c is (that is line 3).
How many bundle sales replace N single sales
bundle sales = ceil(N * Ps / Pb)
Round up for the same reason as in the price guide: 8.1 bundle sales is not enough. This count says what the bundle must sell to match the profit of N single sales. It does not say it will.
A worked example: Kai's bundle (imaginary)
Kai is imaginary. Kai sells a planner template, a habit tracker and a budget sheet. Fees are invented: 10% in percent fees, 0.40 fixed per order, 0.60 other cost per order. Kai assumes the flat amounts apply once per order, and notes it as unchecked.
KAI'S WORKSHEET (imaginary)
Prices: planner 12.00
tracker 9.00
budget 15.00 (sum 36.00)
A. planner: 12 x 0.9 - 1.00 = 9.80
tracker: 9 x 0.9 - 1.00 = 7.10
budget: 15 x 0.9 - 1.00 = 12.50
B. Ps = 29.40 / 3 = 9.80
C. T = 9.80 + 7.10 + 12.50 = 29.40
4. Bundle price: 28.00
D. Pb = 28 x 0.9 - 1.00 = 24.20
E. discount = 1 - 28/36 = 22.2 %
(All figures in this guide were recomputed in Python on 26 October 2026.) Bundle margin is 24.20 / 28.00 = 86.4%, which looks healthy. Look at the price lines instead.
KAI'S PRICE LINES (imaginary)
Line 1: 1.00 / 0.9 = 1.11
Line 2: (9.80+1.00)/0.9 = 12.00
Line 3: (29.40+1.00)/0.9
= 33.78
(6.2 % off 36.00)
Bundle at 28.00: above 2, below 3
The 28.00 bundle is above line 2 and well below line 3. So it beats one average single sale but trails the buyer who would have bought everything. The break-even share is c* = (24.20 − 9.80) / (29.40 − 9.80) = 14.40 / 19.60 = 0.735, about 73%.
Notice how little discount line 3 allows: 6.2% off the sum, even though the margin is 86%. Margin is not room to discount. That gap is the reason I'd run the lines before picking a number.
Now two guesses for c:
- Kai guesses 20% of bundle buyers would have bought all three. Change per bundle sale: 0.8 × 14.40 + 0.2 × (−5.20) = 11.52 − 1.04 = +10.48.
- Kai guesses 30%. Change: 0.7 × 14.40 + 0.3 × (−5.20) = 10.08 − 1.56 = +8.52.
Both are guesses, not data. To replace 20 single sales (20 × 9.80 = 196.00 profit), Kai needs ceil(196.00 / 24.20) = ceil(8.10) = 9 bundle sales. Nine sales give 217.80; eight give 193.60, short.
Kai at other prices
The same bundle at other prices:
| Bundle price | Discount | Pb | c* |
|---|---|---|---|
| 32.00 | 11.1% | 27.80 | 92% |
| 28.00 | 22.2% | 24.20 | 73% |
| 22.00 | 38.9% | 18.80 | 46% |
| 20.00 | 44.4% | 17.00 | 37% |
The deeper the discount, the less would-have-anyway share Kai can tolerate.
When the bundle is a bad idea: Mo's case (imaginary)
Mo is imaginary. Mo sells three cheaper items: 6.00, 5.00 and 4.00, a sum of 15.00. Invented fees: 15% in percent fees, 0.50 fixed, 0.30 other cost per order, applied once per order.
MO'S WORKSHEET (imaginary)
A. 6.00 x 0.85 - 0.80 = 4.30
5.00 x 0.85 - 0.80 = 3.45
4.00 x 0.85 - 0.80 = 2.60
B. Ps = 10.35 / 3 = 3.45
C. T = 10.35
4. Bundle price: 7.50 (50 % off)
D. Pb = 7.5 x 0.85 - 0.80 = 5.575
Line 2: 5.00 Line 3: 13.12
c* = (5.575-3.45)/(10.35-3.45)
= 0.308 (about 31 %)
Mo likes the round half-price offer. It clears line 1 (0.94) and line 2 (5.00), but c* is only about 31%. Mo's imaginary records suggest 60% of bundle buyers would have bought all three. Change per bundle sale:
0.4 x (5.575 - 3.45)
+ 0.6 x (5.575 - 10.35)
= 0.85 - 2.865 = -2.015
That is a loss of about 2.02 per bundle sale under Mo's own guess. Replacing 20 singles takes ceil(69.00 / 5.575) = 13 bundle sales, which looks fine until you notice they would come from the same pool of buyers.
On paper Mo's options are a price nearer line 3 (13.12, or 12.5% off), a two-item bundle, or no bundle. Even at 12.50 (16.7% off) the profit is 9.825, still 0.525 short of three separate sales.
The deeper mistake: a discount deeper than the margin allows
Suppose Mo prices the bundle at 4.00, 73% off 15.00. Profit: 4.00 × 0.85 − 0.80 = 2.60, less than the average single profit of 3.45. Even a buyer who would have bought one item now leaves Mo with less; it clears line 1 and fails line 2. No guess about buyers rescues it.
A 72% margin on a 6.00 item (4.30 / 6.00) does not mean you can discount that far, because the flat fees stay put while the price falls.
The try-or-not rule
This is my judgement, not a benchmark. The thresholds are arbitrary; the value is in writing them down before you look at results.
TRY A BUNDLE ONLY IF ALL THREE:
1. Bundle price >= Line 2
(Pb is larger than Ps).
2. My honest guess for c is no more
than half of c*.
3. I can write one sentence on who
buys the bundle that would not
have bought a single item.
Check 3 matters most to me. If you can't name a buyer who would not otherwise have bought, the bundle is only a discount for people who were going to pay full price. The sentence can be a guess.
Run the rule on both sellers:
- Kai. Check 1 passes (28.00 against 12.00). Check 2: the guess is 30%, half of c* is about 37%, so it passes. Check 3 is for Kai to write. Try it.
- Mo. Check 1 passes. Check 2 fails: the guess is 60%, half of c* is about 15%. Don't try 7.50; re-run at a higher price, or skip.
"Try it" earns you a test, not a conclusion.
Good, better, best: the layout
A tier layout puts three versions of an offer side by side, each with a higher price and more contents. I can offer a layout and its arithmetic, not a promise that buyers will pick the middle one; how buyers choose among tiers is a hypothesis to test.
What each tier must contain
- Good: one complete outcome. It must work on its own: a buyer who picks only this tier can finish the job it promises. A crippled version of the next tier is, in my judgement, a trap.
- Better: Good plus the next thing a buyer needs, such as a tracker for a planner. Name the extra in one line a stranger can read.
- Best: everything, plus a finishing piece such as a quick-start page. If you have nothing to add, Best can simply be "all of it". Don't invent filler.
Each tier needs a one-line "this is for someone who...". If you can't write it, merge the tier into its neighbour. Check each step in profit, not only in price.
The Tier Worksheet
TIER WORKSHEET date: ____
Percent / fixed / other as above
GOOD BETTER BEST
Contents: ____ ____ ____
Sum of single prices:
____ ____ ____
Tier price: ____ ____ ____
Profit = price x (1 - pct/100)
- fixed - other
____ ____ ____
Step in price: ____ ____
Step in profit: ____ ____
Guessed mix of buyers (adds to 100):
___ % ___ % ___ %
Blended profit per sale
= sum of (mix x profit)
= $ ____
Compare: profit if only your
current single offer is sold
= $ ____
Each tier vs line 3: ok / below
Kai's tiers (imaginary)
Kai builds tiers from the same three products, with the same invented fees: Good is the budget sheet at 15.00. Better adds the planner (sum 27.00) at 24.00. Best is all three (sum 36.00) at 30.00.
KAI'S TIERS (imaginary)
GOOD 15.00 x 0.9 - 1.00 = 12.50
BETTER 24.00 x 0.9 - 1.00 = 20.60
BEST 30.00 x 0.9 - 1.00 = 26.00
Step Good->Better: price +9.00
profit +8.10
Step Better->Best: price +6.00
profit +5.40
Discounts vs sum: 0 / 11.1 / 16.7 %
Each step keeps most of the added price as profit (8.10 of 9.00, 5.40 of 6.00), because the flat amounts are paid once. Now the guess: suppose Kai's mix of 100 tier buyers is 50 Good, 35 Better, 15 Best.
Blended = 0.50 x 12.50
+ 0.35 x 20.60
+ 0.15 x 26.00
= 6.25 + 7.21 + 3.90
= 17.36 per sale
Compare: budget sheet alone 12.50
On a 70 / 25 / 5 mix it is 8.75 + 5.15 + 1.30 = 15.20. Both beat 12.50 per sale, but only if the same number of people buy, and only if few Better and Best buyers would have bought those items separately at full price.
Better (24.00) and Best (30.00) are discounted bundles (11.1% and 16.7% off), and both sit below line 3 (25.89 and 33.78). So the would-have-anyway share c applies to each tier just as it does to a bundle. Run the Bundle Worksheet on each tier before you trust the blended figure.
A tier layout could change how many buy, and the worksheet cannot see that. If every buyer chose Good, tiers would equal the single offer. Which mix you get is the unknown.
From arithmetic to a test
The sums tell you what a change would leave if the guess were right. They don't tell you if the guess is right. For that, use the one-change test from Why Your Shop Isn't Growing: write the hypothesis, change one thing, fix the dates, name one measure and write down the result that would change your mind.
Three hypotheses, none with a source:
- H1: "Buyers who find one product add the others when offered a set." Measure: orders per visit, and the share of orders that contain the bundle.
- H2: "Only a small share of my buyers already buy several of these items." Measure: count past orders with two or more of them. That is your evidence for c.
- H3: "A middle tier is picked more than the other two." Measure: the split of tier orders over the test window.
Kai's test: add the bundle listing only, leave single prices alone, measure profit per 1,000 visits over fixed dates, and write down in advance that a lower figure than the same length of time before would change their mind.
A lower result may be the bundle taking from your own singles, and a flat one is no proof either way. With very few visits a test can't tell much; mark the result inconclusive.
Change tiers and bundle in separate tests. Two changes at once spoil the reading.
Common mistakes
| Mistake | Why it hurts | Fix |
|---|---|---|
| Discount deeper than line 2 allows | A one-item buyer leaves less than a single sale would | Price at or above the average single price |
| Reading margin as room to discount | Flat fees stay put, so profit falls faster than margin suggests | Run the three lines first |
| Assuming the flat fee is charged once | Your platform may charge per item | Check the fee page and one real payout |
| Ignoring c | Buyers who'd have bought everything lose you money | Count past orders with two or more items |
| Crippled Good | The first tier looks like a trap | Make Good a complete outcome |
| Changing bundle, tiers and prices together | You can't tell which change did what | One change per test |
Your next step
Fill in the Bundle Worksheet for one set of three products today, with your fee page open. If the rule says no, you've saved an evening; if it says try, write the test before you build anything. If the product pages may not be ready, Audit Your Product Page in 30 Minutes is the cheaper first check, and Write Your First Product Page in Seven Blocks helps with a bundle page.
None of this needs a new tool.
Sources and checks
How we checked. No external source was used: no platform page, statistic or law is cited, and no fee number is quoted. Every figure is an invented example or arithmetic, recomputed with Python on 26 October 2026 (script and output saved in the raw notes folder for this article).
Kai: bundle at 28.00 leaves 24.20 against 29.40 for three separate sales; lines 1.11, 12.00, 33.78; c* 0.735; 9 bundle sales replace 20 singles; tiers 12.50, 20.60, 26.00, blended 17.36 and 15.20. Mo: bundle at 7.50 leaves 5.575 against 10.35; c* 0.308; change −2.015 at c 60%; 13 bundle sales.
Opinion, not fact. The three price lines as a way of framing, the half-of-c* threshold, the contents of each tier, and the order of work are my judgement. The would-have-anyway model (everyone else buys one item) is a simplification.
Unverified. Whether your platform charges its flat fee once per order or per item, whether it allows bundles or tiers in the form described, and everything about how buyers react. Nothing here has been tested in a real shop.
What this guide does not claim. That bundles or tiers sell more, that any discount, fee or margin is normal, that Kai or Mo existed, or that following it will produce profit.