Every creator gets told the same thing about tiers. Offer three. Make the middle one the obvious pick. Add an expensive tier nobody buys so the middle looks reasonable. It spread because there's research behind it, and because it's easy to copy off someone else's pricing page.
The research says something much narrower than the advice does. That helps explain why so many three-tier tables end up converting about as well as the single price they replaced. Worse, the tier that looks most impressive on the page is often the one that produces your least satisfied customers.
The short answer first. Two or three tiers, separated by one thing your buyer already has an opinion about, which for most creators means access rather than volume of content. A course or membership priced on how many lessons or posts you hand over gives the reader a question nobody can answer, and they answer it by picking the cheapest option or closing the tab.
The decoy tier is the weakest idea in the playbook
The "add a tier nobody buys" trick comes from the attraction effect, first shown in a 1982 experiment by Huber, Payne and Puto where adding a clearly worse third option made one of the original two more popular. The finding is real. It's also one of the shakiest things anyone has built a pricing page on.
Yang and Lynn tested it at scale: 91 attempts across 23 product classes and 73 decoyed choice sets returned 11 reliable effects, far fewer than their statistical power predicted. Effects shrank when options were described in meaningful words or shown as pictures rather than abstract numbers, and with realistic stimuli they turned up at chance levels. Frederick, Lee and Baskin hit the same boundary: the effect is largely restricted to numerical, stylized product descriptions and does not typically occur when people actually experience the product, like tasting a drink or seeing the hotel room.
Your buyer is in the condition where it stops working. They've watched you teach and formed a view of what your time is worth long before reaching the page. A padded top tier is something you'll have to deliver if anyone buys it, and until then it's just more text on the page.
What holds up is how hard you make the decision
Iyengar and Lepper set out a tasting table with either 6 or 24 jams, and among shoppers who stopped, roughly 30% bought from the six-jam table against 3% from the 24-jam table. What usually gets left out is that the big table pulled more people in, stopping 60% of 242 passers-by versus 40% of 260 at the small one. Variety got attention, but it lost the sale.
Even that isn't a law. The biggest review of this literature, covering 99 observations from 53 studies and 7,202 participants, found the mean effect of assortment size comes out nonsignificant when moderating conditions are left out. Put the moderators back and a clear effect appears, driven by choice set complexity, decision task difficulty, preference uncertainty, and whether the person is trying to minimize the effort of deciding.
That's more useful than "offer three." Two tiers or five, the number is rarely what decides it. A buyer has to be able to tell them apart and judge which one fits. A page offering 40 lessons, 60 lessons or 85 lessons is textbook preference uncertainty, because nobody browsing knows how many lessons their problem requires.
One price first, then the tier your buyers asked for
If this is your first product, a single price will usually beat a tier table. Tiers let different buyers sort themselves, so you need to know your audience splits into groups that want different things. Before you have buyers, you're inventing that split. Launch at one price and watch what people ask before paying. When people keep asking for feedback, or a deadline, or faster answers, that's your second tier, described by the person who'd buy it. On Sherpo that second tier is another pricing plan on the same product rather than a second product, so it takes minutes to add once the request keeps arriving.
Pick one dimension your buyer already has an opinion about
Your tiers should differ along something a reader can place themselves on right away. Access to you is the strongest one for most creators, because everyone knows whether they want feedback or prefer to work alone. Pace works too (self-paced, or deadlines with a group moving together), and so do accountability and the right to use the material with a team.
People hold those preferences before they arrive. Someone who has abandoned three self-paced courses knows they need a deadline. They can choose without stopping to think, which is the whole job of a tier table. It's the same discipline as working out what your audience is actually trying to get done instead of guessing at features.
Then name the tiers after the difference, since 3 months, Lifetime, Lifetime plus AI access does more work than Bronze, Silver and Gold. For example, you could structure a $399 course like this: three months of access at $199, self-paced with lifetime access at $399, lifetime plus MCP access at $799. It is the same course in all three, if you don't want to limit some specific lessons or files to a certain tier. What changes is the access.
The most loaded tier is the one most likely to disappoint
Research on feature fatigue found 66% of people chose the model with the most features when judging it in a store, while 56% preferred a simpler model after actually using one, and the same authors' model showed that maximizing the appeal of the initial purchase calls for more features than is good for customer lifetime value.
For a course, that shows up as refunds, unused access, and a top-tier buyer who paid for twelve modules, finished two, and never came back, which is expensive in a business where trust is the real asset.
What a higher tier should actually contain
The top tier still needs something in it. Whatever goes in there should cost you almost nothing per extra buyer and be obviously worth paying for. A fixed block of your time qualifies. Three sessions, or sixty minutes a month, booked through your calendar instead of negotiated over email. It's finite, so cap the seats and price it accordingly.
AI access to your material is the bonus almost nobody is offering yet. A Sherpo product can expose an MCP, so a higher tier ships your course as something the buyer's own AI assistant can query while they work. It's material you already wrote, in a form that answers questions at 11pm, and it can even include your exclusive skills or prompts, based on the teaching material.
Ongoing access works the same way. Include your paywalled posts in the top tier and a one-off purchase turns into a relationship. Add a credit toward whatever you sell next, as a coupon with its own link, and you learn later which tier produced repeat buyers. Put top-tier buyers into your affiliate program and the people most likely to recommend you get paid for it. Quizzes and a certificate cost nothing per buyer and give the tier a finish line. A buyer can put a number on any of these. Twelve more modules, they can't.
Change how a tier is paid before you cut what it costs
When the top tier stalls, look at the payment shape first. Stripe's test across more than 150,000 checkout sessions found buy now, pay later raised revenue by up to 14%, with over two-thirds of that volume net-new and the largest gains on orders between $500 and $1,500, exactly where a serious creator program sits.
On Sherpo, one product carries multiple pricing plans at once: a one-off payment, a recurring subscription, or a subscription with an expiry date that ends rather than renewing, which is what a twelve-week cohort actually is. Renewal length is yours to set, down to odd cycles like every three weeks, and course content is gated per plan, so a higher tier really does unlock more without you maintaining two products. Stripe checkout and BNPL come with it. And while you're picking numbers, there's a reason prices ending in 9 keep working.
Let the tiers report back
There are only a few questions worth asking later. Which tier did the people who finished the course buy? Did the mid tier attract upgraders, or buyers who would have paid full price anyway? Did top-tier buyers renew? Those take minutes when the plan, the purchase, the progress and the renewal are one record, and a weekend of spreadsheets when they sit in four tools, which is the real cost of a fragmented stack.
Most tier tables fail the way most launches do: someone polished the page and never thought hard about the decision behind it, which is the mistake at the heart of why most creators fail at monetization. Give people two or three options that differ in a way they can feel, name the tiers after that difference, and keep the expensive one light on content and heavy on access. Then watch what buyers actually pick.
Ready to build tiers people can actually choose between? Get started with Sherpo and run multiple plans, gated content, MCP access and renewals on a single product.
Giacomo Di Pinto
Oct 5, 2026
7m reading time
Read more articles