Publish a benchmark report from your own data

A benchmark is only useful if it measures something readers already track and worry about. Kite starts by looking for those metrics in the questions people ask. It checks search demand for phrases such as "average churn rate for SaaS" or "good email open rate," reads the discussions on sites such as Reddit, X, and Hacker News, and looks at which benchmarks competitors and industry publications already publish. It also checks how AI answers handle the question today. An answer that hedges, cites an old survey, or gives no number at all points to a gap your data can fill.

Kite then matches those questions to the data you actually hold. For a help desk product, that might be first response time, time to resolve a ticket, and tickets handled per agent. For each metric, Kite proposes the peer groups that make the comparison meaningful, such as company size, industry, region, or plan type. A reader at a 20-person ecommerce company does not want to compare themselves with a 5,000-person bank, so the peer groups matter as much as the metrics.

You get a short plan to approve: the metrics, the peer groups, the time period, and the three to five headline findings Kite expects to test. Agreeing on this first keeps the data pull focused on the numbers that will matter.

Kite works from the data you choose to share. It can read events and funnels from a connected product analytics tool such as PostHog, Mixpanel, or Amplitude. It can read payments and subscriptions from Stripe, or deals and contacts from HubSpot or Salesforce. It can also read a Google Sheet or Airtable base where your team keeps the numbers. If your data lives in a warehouse or an internal database, export the aggregated table as a CSV and share it with Kite in Slack.

A benchmark publishes group figures, never a single customer's numbers. Kite calculates medians, ranges, and percentiles for each peer group and reports only those. You set the rules for what can be published, such as "no group with fewer than 25 companies" or "leave out any customer who has opted out of aggregate reporting." When a group falls below your threshold, Kite merges it with a similar group or leaves it out, and tells you which groups it changed. Kite saves those rules in your team's knowledge base, so the next edition follows them without being reminded.

If some customers have contracts that limit how their data is used, remove them from the export before you share it. Kite works with what you give it.

Kite calculates each metric for every peer group, then looks for what's worth saying. A good benchmark finding states the typical value, the comparison that gives it meaning, and the size of the group behind it. For example: "The median support team at companies with 50 to 200 employees sent its first reply in 3 hours and 40 minutes, across 1,140 companies from January to June 2026." A writer can lift that sentence straight into an article, and a reader knows exactly what it describes.

Kite prefers medians and ranges to averages, because one outlier can drag an average far from what a typical company sees. It reports the spread as well as the middle, such as the 25th and 75th percentiles, so readers can tell whether they're merely below average or far behind. Where two sources disagree, such as revenue figures from Stripe and from your CRM, Kite names the conflict instead of silently picking one.

The analysis often turns up a surprise, such as smaller teams replying faster than larger ones. When it does, Kite points it out and suggests making it a headline, because the surprising findings are the ones people share.

Kite writes the report in your brand voice. A typical benchmark opens with the headline findings, then gives each metric its own section with a chart and a table of values by peer group. A short section explains what the numbers mean for the reader and what a company behind its peers might change. A methodology section closes the report. It says which customers are included, the time period, how each metric is calculated, the smallest group size reported, and what was left out.

Each section opens with a heading that matches the question a reader would ask, such as "What is a good first response time for a small support team?", followed by the answer in its first sentence. That structure lets a reader jump to their own row, and it lets an answer engine quote the section without needing the rest of the page.

Kite designs the charts in your brand colors, fonts, and logo. Every chart and table carries a line naming the dataset and time period it comes from. You review the draft and ask for changes in plain language, such as "lead with the resolution time finding" or "drop the region breakdown."

A benchmark built to be cited should be open to everyone, because journalists, other writers, and AI answers can't quote numbers hidden behind a form. Kite publishes the full report on an open page. If you also want to collect leads, keep a downloadable version with the complete data tables behind a form, as described in Produce a gated research report.

Kite can publish the report in two ways. It can add it to your website as a page or article, which keeps the citations and links pointing at your own domain. On a website Kite hosts, it publishes directly and adds the structured data and llms.txt entry that help AI systems understand the page. If your site runs on Webflow, WordPress, Contentful, Sanity, or Storyblok, Kite publishes through your connected content system.

Kite can also publish the report as a hosted report page with charts and tables, which opens from a shared link without signing in. Its charts and tables load without JavaScript, so a crawler that doesn't run scripts can still read every number. For more on hosted report pages, see Publish a shareable report page.

Kite asks for your approval before anything goes live. You approve with the Approve or Publish button in the Kite web app or in Slack, or by replying in the conversation.

A benchmark earns citations when the right writers see it in the first few weeks. Kite finds the journalists, newsletter writers, and bloggers who cover your market and reads what they've written recently. It then drafts a pitch for each one that leads with the finding closest to their beat. A reporter who covers customer service gets the response time finding, while a reporter who covers small business gets the breakdown for companies under 50 people. For more on how Kite researches reporters and personalizes pitches, see Run a media outreach campaign.

Kite also turns the findings into social posts, usually one finding per post with its chart, and suggests timing for the series. Once you authorize it, Kite publishes them through your connected LinkedIn and other social accounts. It can also write the announcement email to your customers and your list, many of whom will want to see where they rank.

After launch, Kite measures the report by readers and citations. On a website Kite hosts, its built-in analytics show how many people visited the report and where they came from. Sources include search, social, email, and AI assistants such as ChatGPT and Perplexity.

For AI answers, Kite tracks the questions your benchmark was built to answer, such as "What is a good first response time for a support team?", across ChatGPT, Gemini, Perplexity, and Claude. It records whether each answer mentions your company and whether it cites your report page, then checks again as time passes. Add Kite to Slack and it can also keep a standing watch for new pages that mention your report or your company, and tell you when one appears.

A benchmark gains authority each time it's updated with the same method. You can ask Kite to rerun the analysis quarterly or once a year. Each run reads the latest data, applies the same metrics, peer groups, and privacy rules, and shows what changed since the last edition. A hosted report page keeps the same link when Kite updates it, so every citation from the last edition still points to the current numbers. The change itself, such as "first response times fell 18% in a year," often becomes the next headline.