Discovering Your Brand from the Web

Most brand identity content doesn’t live on the homepage. It lives on the About, Values, Sustainability, Mission, and Press pages — and the same is true for your competitors. The new Discover from the web workflow in the Brand Analysis tab finds that content automatically and turns it into structured brand traits you can review and refine.

1. What It Does

When you start a web discovery run, Markolé:

Searches the web for each entity you specify, using Google Search grounding.
Crawls up to 75 URLs across a depth of 3, following only the most relevant child links on each page.
Extracts brand-identity content — mission, values, tone, positioning, archetype signals — from every page it visits.
Runs every chunk through the same trait and topic extraction prompts used in the guided Brand Interview, so the output drops cleanly into your brand’s evidence pool.
Streams progress live: a 3-line sanitized preview, a Sub-Category bubble chart that updates as traits land, and a running count of URLs analyzed.

2. Starting a Run

Open the Brand Analysis tab inside your brand.
Find the Discover from the web card at the top of the page.
Make sure your brand has a name and an industry set — these are required.
Optionally add up to three competitors by name. Markolé will search and crawl them in parallel with your brand.
Click Start discovery.

A run takes a few minutes depending on how many URLs are reachable. You can leave the page and come back — the run continues in the background, and progress is restored when you return.

3. What You See During a Run

Live preview — a fixed 3-line area showing the latest content snippets the crawler is processing, sanitized for readability.
URLs analyzed counter — the running total of pages successfully processed.
Sub-Category bubble chart — updates in real time as new traits are extracted, giving you an immediate sense of which categories of brand identity are populating.

The panel is designed not to flicker — preview and chart update smoothly while the crawler works.

4. Trait Provenance

Every trait extracted from the web lands in your brand’s evidence log with full provenance:

Source URL — the exact page the trait came from.
Source type — web_discovery, distinct from interview-sourced or document-sourced evidence.
Depth — how deep in the crawl tree the page sat (1 = seed, 2/3 = followed links).
Entity role and name — your brand or which competitor.
Run ID — the discovery run that produced it.

Competitor evidence is merged into the same DNA pool so you can see overlap and contrast, but the audit trail is always there if you need to filter or attribute later.

5. Brand URLs and Seeds

If you’ve added URLs to your brand under Brand URLs, those URLs are automatically seeded as depth-1 sources for your brand during a discovery run. You don’t need to analyze them separately — web discovery picks them up.

To edit your brand’s URLs, click Edit on the Brand URLs card.

6. Re-Running

You can re-run discovery any time:

Use the Re-run button when you want a fresh crawl (for example, after a competitor changes their site).
Results are appended to your evidence pool — re-running does not erase previous traits.

Web discovery is designed to be safe to run repeatedly. Only one run is active at a time per brand.

7. Tips for Best Results

Pick competitors carefully — choose brands that genuinely overlap with your positioning. Random competitors add noise.
Set the industry field accurately — it improves the relevance of the search step.
Combine with the guided Brand Interview — web discovery gives you breadth, the interview gives you depth. The two evidence streams complement each other.
Review provenance when in doubt — if a trait feels surprising, check its source URL before locking it into your brand.