Quick Start
Continue with the brand created during signup, review buyer questions, read your first scores and choose one useful gap.
On this page
Signup creates your first brand. Continue with that brand here. You do not need to add it again.
- The brand you created during signup, checked and ready to track
- A prompt set that covers real buyer questions
- A first visibility baseline and one gap to investigate
Before you start
Sign in and select the brand you created during signup in the sidebar. Have its public website, the market you sell in, and a few buyer questions to hand.
Tracking needs an active trial or paid subscription, an active brand and active prompts. Check Settings → Billing if tracking is waiting for a plan. You do not need to connect a CMS, Google Analytics or crawler tracking to get a visibility report. Those connections add publishing, visitor or crawler data later.
Step 1: Check the brand you already created
Open Settings → Brands, then select your brand's row.
| Section | What to check |
|---|---|
| Identity | The public brand name, website and description match what you sell. Use a website that can be opened without signing in. |
| Market and language | The Primary market matches your buyers. Shared question language matches the questions you want to track. Check the answer language of each tracked market. |
| Name variants | Add real alternate spellings or product names that should count as your brand. Avoid names shared with a rival. |
| Tracking | The brand is active and has active prompts. |
Setup suggests brand details, competitors and prompts from your website. Review them, especially if your site covers several products or markets. If setup could not read the site, correct the description and questions yourself.
Step 2: Check your prompts
Open Prompts from the sidebar. Start with a small set of category and buyer discovery questions where AI has to choose which brands to recommend.
Category
What are the best running shoes for marathon training?
Buyer need
Which running shoes suit a first marathon on a budget?
Requirement
Which running shoes offer good support for flat feet?
Too broad
best products
A question such as “How does Nike compare with Adidas?” answers a different research need. Naming the brands gives them a place in the question, so its score is not a clean measure of discovery. Trakkr can collect and score branded questions too; the tracking collection asks for ranked brand recommendations, so it is not a full conversational review of your reputation. Keep those questions in a separate tag and use Perception for how AI describes you.
To add questions, choose Add → Add prompts. Add → Suggest prompts helps find gaps. Use the On toggle to deactivate irrelevant questions while keeping their history. Growth and Scale normally include 50 active prompts per brand; the page shows your account's allowance, including add-ons or custom limits.
For a complete selection process, read Choose useful prompts. Keep the starting set stable while you establish the baseline.
Step 3: Watch the first scan
Trakkr queues the first report after setup and plan activation allow tracking. Open the Dashboard for this brand. Its First scan panel shows progress and names a blocker when it can, such as no active prompts or a paused brand. You can close the tab while the scan continues.
Allow a few minutes, and follow any message the panel shows. If it stays stuck, check Troubleshooting. A report in progress is not a score of zero.
Step 4: Read your baseline
Read the date and scope beside the numbers before comparing them.
| Metric | What it tells you |
|---|---|
| Visibility | How prominently your brand ranks in successful tracked answers. Higher answer positions earn more points. |
| Mentions | The count of brand appearances in the stated period. |
| Rank | Your position among the brands observed in the comparison. |
| Citations | Source URLs returned by the citation platforms. Check the label and scope to see which sources the count includes. |
| AI Traffic | AI-referred sessions from a connected analytics property. |
| Conversations | Requests identified as conversation fetches by connected crawler tracking. They do not prove your page was cited. |
AI Traffic and Conversations need their own connections. Missing connection data is not evidence that nobody visited.
Compare the same prompts, market, model and period. Open a low-scoring prompt and inspect the per-model results before making a site change. A measured 0 means successful answers earned no top-ten visibility points. A missing score means no usable measurement for that slice. Zero and missing scores explains the difference.
Step 5: Choose one gap
Pick one buyer question that matters to your business and where a competitor appears ahead of you. A low whole-brand score is a starting signal, not a brief to rewrite the whole site. Leave branded comparison questions out of this choice, because naming a brand in the question changes the test.
Open the prompt and note which competitors rank. In Citations, find the question under Queries and inspect the cited pages. A page marked Not checked is unknown, not a confirmed gap. Then match the evidence to the smallest useful change.
| Evidence | A useful first task |
|---|---|
| Your relevant page cannot be reached, or its text is unavailable to the tested fetch | Open the URL in Pages and fix access or rendering before changing the copy |
| Your page is readable but lacks the answer buyers need | Update that page with a direct answer and sources you can verify |
| No page fits the buyer question | Prepare one focused article in Content and review it before publishing |
| An independent page names rivals but omits you | Check whether inclusion is realistic, then review an Outreach draft |
| A model states something wrong or out of date | Check the claim in Accuracy or Perception, then decide where corrected facts belong |
Not measured means a check or connection is missing. It is not a failed stage. A citation, crawl or competitor ranking does not prove why AI skipped you, so read the evidence before you commit to the work.
Look for matching work in Actions. When What we'll measure appears, check the measure, window and baseline before confirming; the plan is frozen once started. If you do the work outside Trakkr, note the prompt, market, model, target URL and start date, and keep that prompt group stable so a later score compares the same thing. When the window closes, read Results. No movement and Couldn't measure are useful outcomes too.
Common questions
What if I cannot find the brand from signup?
Check the sidebar selector and Settings → Brands first. If you left setup before it completed, return to setup. Avoid making a duplicate of a brand already listed in your account.
Do I have to write prompts myself?
No. Setup suggests a starting set. Review it against what your customers ask, and replace questions that do not fit your category or market.
What if the first scan finds no mentions?
First confirm it completed with usable answers. If it did, no mentions is a real baseline. Read the competing brands and cited pages for one important question, then choose a gap you can address.
How long until I can compare a trend?
You need more than one report. The Dashboard offers longer ranges as history becomes available. Compare dates and coverage, and give a content change time to be discovered before judging it.