How Big Should a Product Launch Actually Be?

How big should a product launch be?

Most launches should be small, a few should be medium, and roughly one a year should be large. The mistake is not choosing wrong once. It is having no tiers at all, so every release gets the same announcement, and the announcement that mattered arrives looking identical to eleven that did not.

Tiering solves this. You decide in advance what qualifies as a big launch, what the big treatment involves, and what everything else gets instead. It takes an afternoon and it removes the recurring argument about whether this release deserves a webinar.

Here is the framework we use with B2B software clients, and the reasoning underneath it, which comes from effectiveness research rather than from launch folklore.

Why does launch size matter more than launch quality?

Because attention is the scarce input, and you get very few chances to spend it. Every large launch draws on the same finite willingness of customers, press and your own team to pay attention. Spend it four times a year and each spend is worth a quarter as much.

The scale of the competition for that attention is easy to underestimate. Zylo's 2026 SaaS Management Index reports that the average company manages 305 SaaS applications, and that marketing teams alone use an average of 103 marketing related applications.

Your announcement is competing with 304 other vendors who also have news. A launch that is merely good is invisible in that field. A launch that is genuinely big can be seen, but only if you have not spent the same volume on the previous three.

What decides the tier?

One question, asked honestly: does this change who can buy from us, or only what existing customers can do? A change to who can buy is a market facing event. A change to what customers can do is a product event. They deserve completely different treatments.

That question maps onto a real difference in audience. John Dawes of the Ehrenberg-Bass Institute puts it as the 95:5 rule: up to 95% of firms are not in the market for a given product at any one time. Existing customers and active buyers are the small remainder.

So a product event speaks to people already paying attention, and it needs reach into your base rather than into the market. A market facing event is your chance at the 95%, which is rare and expensive and should not be wasted on a settings improvement.

What are the three tiers?

Tier one is a market event: a new product, a new category position, a new segment you can now serve. It gets a dated moment, a named campaign, a website change, and external outreach. Once or twice a year at most.

Tier two is a significant capability that changes evaluations but not the market: a major integration, a serious new module, a pricing change. It gets a page, an email to the base, a sales enablement update, and a post. A handful a year.

Tier three is everything else: improvements, fixes, smaller features. It gets a changelog entry and, where it matters to a particular segment, an in product note. No email, no campaign, no meeting to discuss whether it needs a campaign.

What does each tier actually cost?

Write the cost down next to each tier, because that is what stops tier inflation. A tier one launch is typically six to ten weeks of work across product marketing, design, web and sales enablement. A tier two is one to two weeks. A tier three is under an hour.

Once the cost is written down, the argument changes character. Nobody wants to spend eight weeks on a settings feature, but plenty of people will argue it deserves a launch when the cost is invisible.

Put the effort estimate in the same document as the criteria, and revisit both once a year. The criteria drift as the company grows, and a tier one at Series A is a tier two at Series C.

What should a tier one launch actually include?

Four things, and website work is usually the long pole. A real page rather than a blog post, because a page persists and a post sinks. A changed homepage, because a launch your homepage does not acknowledge is not a company level event. Sales material that lands before the announcement, not after. And one specific proof point.

Binet and Field's five principles of growth in B2B marketing, published with the B2B Institute using IPA Databank data, name two that matter here directly: invest in share of voice, and maximise mental availability. A tier one launch is one of the few moments a B2B company can genuinely raise its share of voice.

Their fifth principle is to harness the power of emotion, and their report notes that emotions have a much more significant role in B2B than many would believe. A launch is the one moment where a B2B company is permitted to be interesting, and most spend it listing capabilities.

The website side of this is the part that reliably slips, which is why we wrote a checklist for the web work behind a launch.

What does a tier three release need, given it gets no campaign?

A changelog that people can actually subscribe to, and an honest entry for every shipped change. That is the whole mechanism, and it does more for customer trust than a quarterly email that bundles twelve things nobody asked about.

A good changelog also solves a problem tiering creates. If small releases get no announcement, customers can feel the product has gone quiet. A public, frequently updated changelog is the visible evidence that it has not.

Make it a real page on your own domain rather than a tool on a subdomain, so it accumulates search value and answer engine visibility, and so a prospect evaluating you can see momentum. The build details are in how to build a changelog page.

How does AI search change launch planning?

It reduces the value of the press hit and increases the value of the durable page. Pew Research Center tracked 900 United States adults across 68,879 Google searches in March 2025 and found that when an AI summary appeared, users clicked a result in 8% of visits, against 15% when no summary appeared.

Pew also found users ended their session entirely 26% of the time after a page with an AI summary, against 16% without. Coverage that would once have sent traffic increasingly gets summarised instead.

The practical consequence is to move budget from the moment to the artefact. A launch page that clearly states what the product is, who it is for, and what it replaces is readable by answer engines for years. A press release is readable for a week.

How many tier one launches can a company sustain?

Fewer than the roadmap wants. Our rule of thumb is one per year for companies under roughly fifty people and two for larger ones, because the constraint is not the product, it is the organisational capacity to do the surrounding work properly.

Companies that attempt four discover a pattern. The first is excellent, the second is good, the third is late, and the fourth is a blog post with a launch label on it. The tier degrades silently until it means nothing.

If your roadmap genuinely contains three market changing things in one year, that is a real problem worth escalating, not a scheduling puzzle. It usually means the company is trying to enter more markets than it can support, which is the trade covered in expanding your market versus going deeper.

Who decides the tier, and when?

One named person, at the start of the quarter, using the written criteria. Not a committee, and not at the end when marketing is asked what they can do in two weeks. The tier decision is a planning input, because tier one work has a lead time that late decisions cannot accommodate.

Give that person the right to say tier three to something the product team is proud of. That is the entire value of the role, and a tiering system where nobody ever says no is just a naming convention.

Write the decision down where engineering can see it, with the reason. Most resentment about launch decisions comes from the decision being invisible rather than from the decision itself.

How do you know the system is working?

Two signals. The number of tier one launches per year stays where you set it, and the changelog stays busy. If tier ones creep upward, the criteria are too loose. If the changelog goes quiet, small releases are being held back waiting for a launch that will not come.

The second signal is qualitative and more telling: whether your sales team can name this year's big launch without checking. If they cannot, you did not have one, whatever the calendar says.

We build the pages, sites and content systems that launches run through, so we usually see the tiering decision arrive too late to serve it well. If you want help setting the criteria before the next planning cycle, we are happy to work through it with you. You can find us at phoenix.studio.