Getting started

Website setup is a form with six steps. Fond fills in most of it from your own site. The most important step is the last, where you decide which questions Fond tracks.

Before you startCopy Anchor Link

• An account and a plan. A new organization chooses a plan before it can reach the dashboard. Your plan sets how many questions you can track, which assistants your runs can use, and how many runs you get. It does not set the interval between scheduled runs, so an upgrade does not make them more frequent.
• The URL of the store that you want Fond to track. You set up one website at a time.
• Three to five competitors. Fond suggests some from your site, and you can replace any of them.

Adding a website and what Fond readsCopy Anchor Link

Enter your URL and click Detect. Fond reads your site in a few seconds.

• Fond reads up to 20 pages, starting from your homepage and your sitemap. It skips account, policy and other utility pages.
• From those pages, Fond writes a brand profile of what you sell, your values, your markets and your products.
• Fond reads product data, your social profiles and your blog link from the structured markup on your pages.
• Fond proposes the product categories that match what it found.

Then you confirm the store name, the primary contact and the product categories. The categories decide which questions Fond generates, so choose the ones that describe what you sell.

If Fond cannot read your siteCopy Anchor Link

Fond tells you which of two problems occurred:

• Fond could not read your site at all, because of a bot wall, a redirect loop or a timeout. Fond prefills nothing.
• Fond read your site but could not summarise it. Fond prefills the product list from your markup, but not the written profile.

In both cases, type in the blank fields yourself. Fond writes your questions from them, so they are worth a few minutes.

Brand, audience, competitors and presenceCopy Anchor Link

The four steps after the URL are the profile that Fond writes your questions from. Fond prefills each step where the crawl found something.

• The Brand Identity & Products step asks for your description, values and primary markets, and your products by revenue and by strategic importance.
• The Audience & Query Signals step asks who your customer is, which shopping questions you want to win and which questions customers ask you. It also asks for site-search terms, converting ad keywords and subjects to avoid.
• The Add Competitors step asks for three to five brands to measure you against. Accept Fond's suggestions, replace them or type your own. For each one, say why it is your closest competitor.
• The Online Presence step asks for your social profiles, follower ranges, subreddits, blog, YouTube channel, press and review platforms. Every field is optional.

The competitors leaderboard lists the brands that AI answers mentioned. A brand that you track appears only in a view whose answers mentioned it. Brands that you did not name can appear too.

Reviewing the generated questionsCopy Anchor Link

The last step shows the questions that Fond wrote for your brand. Take the most time here. Fond measures every rate on the visibility and competitor pages over these questions only.

Each question has three labels that you can change:

• Tier is how hard the question is to win. Safety is closest to your brand and least competitive, then come Strong, Reach and Far Reach. Far Reach is category-level, so track it as a goal. The step has one tab for each tier.
• Type is the kind of question: Brand Reputation, Comparison, Category, Problem-Solving, Local or Brand Identity. Brand Identity and Brand Reputation describe the brand that the question asks about, not a shopper's task.
• Focus tells you whether the question is about a product or about your brand as a whole.

Edit, delete or add questions until they describe your market. The step shows your plan's question limit. If you go over it, the page tells you how many questions to remove before you can save.

A question's history starts when you start to track it. Adding or removing questions later changes what the rates cover. So the visibility trend marks the runs where the question set changed.

The first runCopy Anchor Link

What happens when you save depends on your plan.

• On the free audit plan, Fond starts your one run immediately. It takes you to the results, which fill in as the answers arrive.
• On a plan with ongoing monitoring, the save takes you to the website overview. Fond picks up the site at the hour that it gave the site, usually within a day. To see numbers sooner, click Run Now.

While a run is in flight, the Run button shows its progress, and the Query Results page fills in answer by answer. How Fond works gives the run statuses and what happens to a run that takes longer than three hours.

Until a run finishes, the Visibility page says that there are no completed runs yet.

After that, a scheduled site runs again every other day in the same hour, unless Fond has set another interval. Runs that you start count against the weekly run ceiling of your plan.

What to do nextCopy Anchor Link

• Read how Fond works once, so that you know what the numbers mean.
• Read the Visibility guide before you draw conclusions from the trend. The trend has one point for each run, not for each day, so it shows little until several runs are complete.
• Run the Agent Readiness checks. They are about your site, not about answers, and they start on their own.
• If a figure surprises you, look in the glossary.

TroubleshootingCopy Anchor Link

Detect found nothingCopy Anchor Link

Your site refused the request, redirected in a loop or timed out. Type the fields in yourself. Nothing after the crawl needs it.

You have more questions than your plan allowsCopy Anchor Link

The step gives the number of questions to remove. A removed question leaves future runs, but what past runs measured stays.

Run Now is unavailableCopy Anchor Link

A run for that website is already in flight, and Fond allows one at a time. The button comes back when that run reaches a terminal status.

A run stuck past three hours does not block forever. A REST API request fails a stuck run and starts the new one, unless the stuck run can still get answers.

If Fond refuses the run, read its message. The usual reason is a spent run ceiling, which you clear by waiting or by an upgrade. You cannot clear the other reasons:

• The question set has nothing in it that can run.
• No assistant is enabled for the site.
• Fond's own configuration is incomplete.

The run says "Completed with errors"Copy Anchor Link

At least one request to an assistant failed. This is normal, and Fond does not count the failures against you. How Fond works gives how each figure handles them. To see which requests failed, open the Query Results page. It offers a retry where Fond can run the failure again.

Visibility says there are no completed runs yetCopy Anchor Link

Visibility reads only the runs that it can serve. So this message appears in these cases:

• No run has finished.
• Every answer in the finished run errored.
• The run was marked failed, also after its answers came in, or it is hidden from analytics.

The status of each run on the Query Results page tells you which case it is.

Organic visibility reads n/aCopy Anchor Link

Either every question in the run named your brand, or every answer to a question that did not name you failed. Organic visibility explains both causes and what to add.

What this does not coverCopy Anchor Link

• Connecting GA4 or Cloudflare. Each has its own setup guide.
• Reading your data with an API key. The REST API reference covers this.
• Fixing what the numbers show. Agent Readiness is about what assistants can read from your site. Site Audit is a review of the site itself.