SaaS brand architecture

Almost nobody chooses their brand architecture. They ship a second product on a Tuesday, give it a clever name because the founder liked it, and three years later a VP of marketing is running a twelve month consolidation project to undo that Tuesday. The decision gets made by accident and paid for on purpose.

This page covers the four models, the three questions that should decide which one you run, what each choice does to your search footprint and site structure, and the migration sequence for cleaning up a house of brands you never meant to build.

What are the four brand architecture models?

There are four, and they sit on a spectrum from one name to many. The further right you go, the more marketing budget the structure demands.

Model	How it looks	SaaS example	Marketing cost	Best for
Branded house	One name, descriptive product labels	Salesforce Sales Cloud, Service Cloud	One demand engine	Overlapping buyers, suite selling
Sub-brands	Parent plus named tier	HubSpot Marketing Hub, Sales Hub	One engine plus per-hub positioning	Same buyer org, different budget holders
Endorsed brands	Product name plus visible parent	Twilio SendGrid	1.5 engines	Post-acquisition, strong acquired equity
House of brands	Independent names and sites	Zoho's early product sprawl	One engine per brand	Non-overlapping buyers, different markets

The SaaS test for each one is blunt. Branded house: can a single homepage headline describe all your products without becoming meaningless? Sub-brands: do different budget holders inside one company buy different products from you? Endorsed: did you buy the product, and does its name get typed into Google more than yours? House of brands: would a buyer of product A be actively confused or put off by product B?

Most companies fail the last test and build a house of brands anyway.

Which three forces should actually decide the model?

Three questions, answered honestly, settle it in an afternoon. Everything else is decoration.

Do buyers overlap? Not “could they” but do they. Pull your CRM. If more than 40 percent of product B’s pipeline came from accounts that already know product A, you have one market and one brand. Atlassian’s Jira and Confluence buyers overlap almost completely, which is why the branded house held even when the product names stayed distinct.

Are the products bought together? Different from overlap. HubSpot’s Marketing Hub and Sales Hub are often bought in separate cycles by separate budget owners, which is why the Hub sub-brand structure works: it lets each Hub carry its own value proposition while the parent carries the credibility. If products are genuinely bought in one purchase order, stop naming them separately.

Was product two built or acquired? Built products should be born inside the master brand, always. There is no equity to preserve and no reason to fund a second engine. Acquired products carry existing search volume, backlinks, review counts and customer recognition, and destroying that on day one is expensive.

What does each model cost you in search?

A separate brand on a separate domain means building topical authority twice. That is the real bill, and it is bigger than the design invoice.

Your parent domain has a link profile built over years. A new domain starts at zero. You now need a content team, a digital PR motion and a link acquisition budget for a second property, and the two properties cannot pass authority to each other in any meaningful way beyond a footer link. Teams routinely underestimate this by a factor of three when they plan SaaS branding cost.

Subfolder versus subdomain matters less than the internet argues about, but it still matters. A subfolder inherits the parent’s authority more cleanly and is easier to interlink. A subdomain is cleaner operationally when the acquired product runs on a different stack and the engineering team refuses to proxy. Pick the subfolder unless engineering genuinely blocks it.

Then there are the places nobody checks. G2 and Capterra categories: two brands means two listings, two review counts split, and two category rankings that are each weaker than one combined listing would be. If your product has 180 reviews split 120/60, you rank worse in both categories than a single 180-review profile would. Review volume is also a heavy input into which vendors language models name when someone asks for options in a category.

The second homepage question is the one that reveals whether the architecture is real. If the acquired product keeps its own homepage, its own nav, its own pricing page and its own trial flow, you are running two companies. If it becomes a product page under the parent nav, you are running one. There is no honest middle.

How does an acquisition actually get integrated?

Slowly, and in a specific order. Twilio bought SendGrid in February 2019 for about $3 billion in stock. SendGrid kept its name, its docs, its domain and its developer community for years, then gradually appeared as Twilio SendGrid in product naming, then folded further into the Twilio platform narrative. That is the normal arc, not a failure of nerve.

Twilio’s Segment acquisition in 2020 followed a similar pattern with a different ending: Segment’s developer brand was strong enough that it survived as a named product long past the point where a weaker brand would have been absorbed.

What do Atlassian, HubSpot, Freshworks and Zoho teach?

Four companies, four outcomes, and the differences are instructive.

Atlassian kept distinct product names (Jira, Confluence, Bitbucket, Trello) under a strong master brand. It works because the buyer is usually the same engineering or IT org, the products genuinely interoperate, and Atlassian’s marketing engine covers all of them from one content library. Trello, acquired in 2017 for $425M, stayed independent-looking longest because its self-serve audience barely overlapped the enterprise Jira buyer.

HubSpot built the Hub sub-brand system deliberately. Marketing Hub, Sales Hub, Service Hub, Content Hub, Operations Hub. Each Hub has its own positioning and pricing but shares one domain, one content library and one enormous brand. It is the cleanest sub-brand execution in B2B software and it depends on HubSpot’s search footprint being one property rather than five.

Salesforce ran Clouds for years, which is the same structure with worse names, and then acquired brands that never fully absorbed. Slack, bought in 2021 for $27.7B, still runs its own domain, its own brand and its own marketing. That is the right call, because Slack’s brand recognition among end users exceeds Salesforce’s.

Freshworks went the other direction and consolidated. Freshdesk, Freshsales, Freshservice, Freshchat all carry the Fresh prefix, so the architecture reads as one company even though the products serve different buyers. It is a sensible compromise for a company that grew by adding adjacent products.

Zoho accumulated dozens of products and for a long time looked like a house of brands held together only by a shared login. The consolidation toward Zoho One as a single suite narrative was the correction, and it took years.

We budgeted for a rebrand. What we actually bought was a year of nobody launching anything.

When is a second brand genuinely right?

When the buyers do not overlap and one product would damage the other’s credibility. That is nearly the only case.

A security product sold to CISOs and a self-serve design tool sold to freelancers have no shared narrative, no shared content library and no shared sales motion. Running them under one brand makes both harder to explain. Keep them separate and accept the cost of two engines.

The other legitimate case is regulatory or contractual: a product serving a market that requires a separate legal entity, or an acquisition where the purchase agreement constrains the brand for a period. These are rare and they are not strategy, they are constraints.

Everything else is a preference dressed as a strategy. If your second brand exists because the founder liked the name or the acquired team objected to losing theirs, you are funding a demand engine for emotional reasons. Say that out loud in the room and the decision usually resolves itself.

How do you consolidate an accidental house of brands?

Pick the surviving brand on evidence, sequence the migration, and protect the revenue-producing pages until last.

The honest tradeoff: you will lose traffic. Some percentage of the retired brand’s rankings will not come back, and branded search on the dead name decays over roughly eighteen months with no replacement. Teams that promise leadership a clean transfer with no loss are setting up a bad conversation in month four. Promise a dip and a recovery, and show the dip in the plan.

One more failure mode. Consolidation projects stall because nobody owns the sales enablement half. The website moves, the decks do not, and reps keep pitching the old name for a year. Assign an owner for sales collateral on day one, with a deadline earlier than the website launch.

What to do this week

Answer the three questions with data, not opinion. Pull the CRM overlap number, check whether product two was built or bought, and look at whether anyone buys both in one purchase order.

If you are under $50M ARR and you have two brands, start the consolidation case now rather than after the next launch makes it worse. If you are about to launch product two, put it inside the master brand unless you can name the buyer who would be confused. And if you are choosing external help, read SaaS branding agencies and the tradeoffs in branding agency vs in house brand team before you write a brief, because architecture work and identity work are different engagements with different price tags. The wider SaaS branding hub covers where this sits relative to brand strategy, and brand tracking tools for SaaS will tell you whether the consolidation actually moved recognition. For visual reference points, the SaaS branding examples collection shows how these architectures look in practice, and SaaS creative agencies covers the execution side once the structure is settled.