SaaS promotion strategy

A 20 percent discount on a $30,000 contract does not cost you $6,000. It costs $6,000 in year one, the same again at every renewal, and the credibility of your price list with the next three buyers who hear about the deal from a peer. Subscriptions make a concession permanent in a way that one-off products never do.

The offers that genuinely pull a deal forward almost never touch price. Teams reach for the discount because it takes ten seconds to construct, and because nobody has built the alternatives and written them into the price book where a rep can find them under pressure.

The arithmetic that should end most discount conversations

Cut list price by 20 percent and you need 25 percent more customers to land in the same place on revenue. Hold gross profit flat instead, at a typical 80 percent software gross margin, and you need 33 percent more. The second number is the one that matters, and almost nobody runs it before the approval goes through.

Look at the bottom row. A 30 percent discount, which most B2B SaaS companies hand out at least once a quarter to close a forecast, means finding 60 percent more customers just to break even on gross profit. No promotion in the history of software has produced 60 percent incremental volume from a price cut alone.

Then there is the part unique to subscription. In a transactional business the discount ends when the transaction does. In SaaS the discounted rate sits in the contract and carries into renewal, so a 20 percent cut on a $30,000 account is a $6,000 annual drag for as long as that customer stays. Three years in you have given away $18,000 and capped what expansion can contribute to net revenue retention on that account, because your first job at renewal is clawing back to list before you sell anything new. That is the quiet mechanism behind companies who add logos at plan and watch NRR sit below 100 percent anyway.

Usage-priced products have it worse still. At Twilio or Snowflake style pricing, a rate concession applies to every unit the account will ever consume, including all the growth you were counting on to make the land worth it in the first place. Anyone running a land and expand motion should treat a rate discount as a tax on the expansion they have not earned yet.

The five offers that move a deal without moving the price

Each of these costs you something real. The difference is that the cost is service capacity, which you can schedule and cap, rather than price integrity, which you cannot get back.

Offer	What it actually costs you	Best for	Failure mode
Extended trial, 14 extra days	Support and CS time only	Self-serve products where activation is the blocker	Extending for accounts that never activated, which just delays the no
Migration or switch assistance	4 to 20 hours of solutions engineering	Competitive displacement off an incumbent	Unbounded scope, because nobody wrote down where the help stops
Annual prepay incentive, two months free	16.7% of year one ARR, paid for in cash forward	Companies that need working capital more than revenue optics	Monthly list price becomes theatre and everyone negotiates to annual
Usage credits, $500 to $5,000	Cost of goods only, roughly 20% of face value	Consumption-priced products	Credits expire unused and the account never reaches habit
Bundled onboarding or implementation	10 to 40 hours of customer success time	ACV above $25,000 with genuine setup work	You give away a service line you meant to sell separately

Migration assistance is the most underused of the five. If a buyer is sitting on three years of tickets in Zendesk or six years of boards in Asana, the real cost of switching is not your subscription, it is the fortnight of internal disruption they cannot get authorised. Offering 12 hours of a solutions engineer, scoped in writing, with a named person and a start date, removes more friction than 20 percent off ever will. It also costs you maybe $1,800 of loaded engineering time on a $40,000 deal.

Extended trials need a condition attached. Give the extra 14 days only to accounts that have completed a defined activation event: invited a second user, connected a data source, run a workflow end to end. A trial extension for a dormant account is a reporting artefact. The team feels generous, the conversion rate does not move, and the deadline that was doing the actual work of forcing a decision quietly disappears.

Usage credits are the cheapest offer on the list and the most misread. A $2,000 credit on a consumption product costs you about $400 in real terms at software margins, and it gets the account past the point where usage becomes habit. Pair it with an expiry inside 60 days, because an open-ended credit gives nobody a reason to do anything this week.

When is a discount actually defensible?

Three cases, and only three. You are buying a named reference you can use publicly. You are buying cash through multi-year prepay. Or you are pricing a pilot low on the explicit condition of a contracted step-up to full rate in year two.

Each one has a test. If the answer to “what am I buying with this money” is anything other than a logo I can name, cash in the bank now, or a signed ramp schedule, you are not making a deal, you are losing a negotiation. “They asked” is not a reason. “It is the last week of the quarter” is a description of your forecasting problem, not a pricing input.

The reference case is worth a specific note. A discount for a logo has to be written into the order form: the customer agrees to appear on your site, take two reference calls a quarter, and give one quotable line with a number in it. Without that clause you have given away 20 percent in exchange for a verbal promise from someone who will change jobs in 19 months. This matters most in enterprise SaaS marketing, where a single recognisable logo genuinely does open the next six conversations.

The annual discount cap, and who approves the exceptions

Set the cap before the quarter starts, publish it, and make finance the owner rather than sales. A structure that holds up in practice: nothing above 10 percent without a manager, nothing above 20 percent without the revenue leader, nothing above 30 percent ever, and no more than 15 percent of quarterly new ARR carrying any discount at all.

That last constraint does the real work. Individual approval thresholds get worn down one deal at a time, because every request looks reasonable on its own. A quarterly budget forces a choice: this deal or the next one. Reps start bringing non-price offers to the table by week three, which is exactly the behaviour change you were after.

Launch moments that create urgency without a price cut

Urgency in software comes from dates that exist anyway. The strongest and most honest is a published price increase with a grandfather deadline: tell the market on 1 October that prices rise 1 December and anyone who signs before then keeps the current rate for 24 months. That is real, verifiable, and it converts a pipeline of “maybe in Q1” into a dated decision without touching what you charge.

Product Hunt still produces a genuine spike for developer and prosumer tools, and the mechanics reward preparation more than budget. A launch that lands in the daily top five typically means a few thousand site visits, a burst of signups skewed toward tyre-kickers, and a durable backlink. It is a bad channel for a $60,000 ACV compliance product and a good one for anything a single person can adopt on a Tuesday afternoon.

Feature launches are the underrated version. Linear’s changelog and Figma’s release cadence work as promotion because each release gives the market a reason to look again, and because both companies write them for users rather than for press. Pick four dates a year, batch the shippable work into them, and you have four campaigns with no discount and no media spend. The pattern generalises well beyond design tools, and it is one of the more reliable SaaS marketing ideas for teams with more engineering than budget.

Category events give you a third kind of moment. If your buyers all attend one conference, the two weeks around it are the only time of year when your entire market is paying attention to the problem you solve. Publish your original research that week. In vertical SaaS this is close to decisive, because the annual association conference is genuinely where the whole industry looks up.

Referral and partner co-promotions that cost nothing up front

Two-sided referral is the cleanest promotion in software because it only pays out on success. Dropbox made the mechanic famous with storage on both sides, and the modern B2B version usually gives the referrer account credit and the referred account an extended trial or waived onboarding. Set the payout at roughly one month of ACV and cap it, because uncapped referral programmes attract affiliate operators rather than customers.

Integration partnerships return far more per hour spent, and almost nobody runs them properly. Find a product with the same buyer and no overlapping functionality, build or document the integration, then launch it jointly: their changelog, your changelog, both marketplace listings, one shared asset. The shared asset should be a template or a working setup guide, not a webinar. Webinars with partners produce attendance numbers and very little pipeline, because the two audiences turn up expecting different things.

Marketplace placement deserves a line of its own. Listings on the Slack app directory, the Atlassian Marketplace, the HubSpot App Marketplace and Zapier’s app index send buyers who already have a specific job in mind. A well-built Zapier listing has quietly outperformed six figure ad budgets for plenty of mid-market tools, and the whole thing costs engineering time. For teams running SaaS marketing with no budget, this is the first place to look before anything paid.

How to tell whether the promotion did anything

Gross numbers always look good during a promotion. The question is how much of that would have happened anyway, and the only way to answer it is with a baseline and a holdout.

The 90 day cohort check is the one teams skip and the one that changes decisions. A promotion that brings in 30 percent more accounts at 25 percent off, which then churn at twice the baseline rate, has actively destroyed value while every report in the building showed green.

A promotion calendar with no discounts in it

Here is a year that generates four genuine moments and never reduces a list price. It assumes a mid-market B2B product somewhere between $10,000 and $50,000 ACV.

Notice the Q4 line. That single move does what Black Friday pretends to do, with none of the damage, because the deadline is about a rate going up rather than a rate coming down. It also sets you up for the following year rather than borrowing from it.

Sequencing depends on your motion. Self-serve products should run these inside the product, at the paywall and in lifecycle email, where the person already using the thing sees the offer. Sales-led companies run them through reps with documented rules. If you are unsure which set applies to you, the split between product led growth and sales led growth is the deciding variable, and it changes almost every implementation detail below the strategy. Getting the whole promotional calendar to fit a coherent SaaS go to market strategy matters more than the individual offers do.

What to do in the next two weeks

Pull every deal from the last four quarters that closed with a discount above 15 percent. Calculate the total ARR given away, then multiply by average customer lifetime in years. That number, on one slide, ends more internal arguments than any framework.

Then build the offer sheet. Five offers, the cost of each, the approval level, and the exact sentence a rep should say. Put a quarterly discount budget in front of finance and get it approved before the quarter opens rather than defending it deal by deal in week twelve.

The broader mechanics of positioning, channel choice and messaging sit in the wider SaaS marketing playbook, and the tactical execution of individual campaigns is covered in the guide on how to promote a SaaS product. Promotion is the layer on top. It works when the price is right and the positioning is clear, and it papers over neither.