Because a wall of logos answers the wrong question. The buyer is not asking whether you have heard of Salesforce. They are asking whether your product will sync the specific thing they need, in the direction they need, without an engineer. A logo cannot answer that, so they leave and ask someone who can.
This page is usually the last stop before a demo request and it is almost always the least designed page on a B2B site. Somebody built it in an afternoon with a grid of images, and it has never been touched since.
We think it deserves the attention a pricing page gets, for the same reason: it is where a real objection either gets resolved or quietly ends the evaluation.
It is removing a blocker, not generating desire. Nobody buys software because it integrates with things. They refuse to buy software that does not integrate with the one thing their workflow depends on. That makes this a page about reassurance and specificity rather than persuasion.
It has a second job that is easy to miss. It captures search demand from people looking for the combination rather than for you. Somebody searching for how to connect two tools they already use is a better prospect than somebody searching your category, because they have already described their situation.
Design for both. The directory serves people who arrived on your site already. The individual pages serve people who arrive from a search for one specific pairing.
List them all, but do not present them as equal. A buyer who cannot find their tool assumes the answer is no. A buyer who finds their tool listed with a thin description assumes it is real but basic, which is usually accurate and always better than silence.
The way to do this honestly is depth tiers. Native integrations get a full page each. Integrations available through a connector get a shorter page that says so plainly. Anything on the roadmap either gets said out loud with no date, or gets left off entirely. Never imply something exists because a logo appears in a grid.
The dishonest version of this page is common and it costs more than it earns. A prospect who discovers during onboarding that the integration was a Zapier connector rather than a native sync becomes a refund conversation, and often a review.
Five things, and most pages have one. Say what the integration does in a sentence a buyer would recognise. Say which direction data moves. Say what specifically syncs, named as fields or objects rather than as concepts. Say what setting it up requires. Say what it does not do.
That last one is the most valuable and the most avoided. Naming the limitation is what makes the rest of the page believable, and it filters out the customers who would churn. In our work, the pages that answer the awkward question directly outperform the ones that dodge it, because dodging just moves the question to a sales call that may never happen.
Add a screenshot of the actual integration in use. Not a marketing illustration of two logos joined by an arrow. Buyers are checking whether this is real, and a real screenshot settles it faster than any copy.
Every integration gets its own URL. Not a modal, not a filter state, not an accordion. If the content only exists after a click inside a JavaScript component, it cannot be linked, cannot be found in search, and cannot be cited by an AI answer engine.
Beyond that, give the directory a search box and category grouping, because buyers arrive knowing exactly what they want. Categories should match the buyer's mental model, so CRM, support, billing and analytics, rather than your internal API taxonomy.
This is a natural fit for templated pages built from a content collection, since every integration shares a structure. Our guide to programmatic SEO covers how to do that without producing hundreds of thin pages, which is the failure mode.
Because does X work with Y is the most answerable question in software, and models answer it constantly. If your integration pages are specific and readable, you are the source. If they are a logo grid rendered by script, somebody else answers on your behalf, and they may be wrong.
The opportunity is larger than most teams assume. Webflow analyzed the websites of 2,000 companies for the study it published in September 2026 and found the median company appeared in only 16% of the AI answers it would want to be part of, and was cited in those answers just 6% of the time.
And the click economics reward being quoted rather than ranked. The Pew Research Center reported on July 22, 2025 that when a Google AI summary appeared, users clicked a traditional search result in 8% of visits, against 15% when no summary appeared, with in summary link clicks at 1%. Being the source the answer was built from is worth more than it used to be.
Say what the workaround is, in plain terms. Most products connect to far more than they natively support, through a general automation layer or an API. Zapier states on its own apps page that it connects to more than 10,000 apps, and for many buyers that is a perfectly acceptable answer if you tell them.
Give the API its own honest page too. Developers evaluating you want to know about authentication, rate limits and whether there are webhooks, and a link labelled documentation that leads to a marketing page will lose them immediately.
What does not work is a request an integration form with no visible outcome. If you collect those, publish the count or the top requests. A page that shows demand being tracked is credible. One that silently swallows submissions is not.
Logos. An integrations directory is the single most image heavy page on most B2B sites, and it is usually built with a hundred full colour images loaded at once, unsized, above and below the fold alike.
That runs straight into the thresholds. Google's Core Web Vitals guidance sets good at 2.5 seconds for Largest Contentful Paint, 200 milliseconds for Interaction to Next Paint, and 0.1 for Cumulative Layout Shift, measured at the 75th percentile of page loads and segmented across mobile and desktop. Unsized logos are one of the most reliable ways to fail the third one.
The fixes are ordinary. Use modern image formats, set explicit dimensions, lazy load anything below the fold, and consider a single sprite or monochrome treatment for the grid. Our notes on image optimization cover the detail.
We would delete the grid from the top of the page and put a search box there instead. The grid is designed to impress, and buyers are not there to be impressed. They arrive with a specific tool in mind, and every second spent scanning logos is friction.
Then we would rewrite one integration page properly, as a template, before touching the other forty. Teams usually try to improve all of them at once, run out of energy, and ship forty half improved pages. One excellent page tells you what the template should be.
Finally we would link each integration page to the feature it enables, rather than to a generic demo request. Somebody reading about a CRM sync is interested in the CRM workflow, and sending them to a page about that converts better than sending them to a form. Our piece on feature page design covers that destination.
Pick your three most requested integrations and write a real page for each. What it does, which direction data moves, what syncs, what setup takes, and what it does not do. Add one honest screenshot. That is a day of work for the pages that carry the most objections.
Then check that each has its own URL and loads without JavaScript. If the content vanishes with scripts disabled, no search engine or AI answer engine can use it, and the effort is wasted. If you want help templating an integrations directory or auditing the one you have, we are happy to look at it with you at phoenix.studio.
Tell us where you want to go. We'll tell you how we'd get you there.