If you use research results in a content or outreach plan, you need to know which collection produced them and what its dates mean. The research reports run on two separate collections. They hold different material, reach back different distances, and answer different questions. Before you read any result, know which one produced it.
The News Index
The News Index holds pages, podcasts and videos collected from news sources and from more than four million feeds.
Understand its coverage window before you rely on it. The index keeps two monthly tables, so it reaches the current month and the one before. In any thirty days that comes to roughly two million documents.
So it is a current-coverage index and not an archive. It tells you what is being published in a market now. It will not tell you what a publication wrote about your brand two years ago. Save the individual results you need while they are in range; bulk export of the whole index is not included.
The Prospect Index
The Prospect Index holds pages collected from the wider web, not from news feeds. These are the pages carrying placement opportunities: guest contribution pages, resource lists, expert roundups, podcast pages, industry directories, award programs and speaking submissions.
It holds more than 10 million prospect pages and takes in about 10 million strong new ones each month. The filters favor prospects you can use over a bigger raw number.
Nothing is dropped for age, because there is no publication-date condition. A resource page from 2019 that somebody still maintains is worth as much as one published last week.
Two dates, and why the difference matters
Every document carries two dates. Confuse them and you will draw the wrong conclusion about timing, every time.
The indexed date is when the document entered the collection. The published date is when its publisher put it online.
They can sit months apart. A page published in spring can be collected today and show up as newly indexed.
The nightly count treats them differently on purpose.
- A result counts as new when it was indexed inside the 24 hours ending at the run.
- In the News Index it also has to carry a published date inside the last seven days. The index rotates monthly, and without that second condition every routine rotation would read as a wave of fresh coverage.
- The Prospect Index has no publication-date condition, because an old resource page turning up for the first time is a genuinely new opportunity.
What follows from that: a link or mention in this morning's report is no evidence it was published last night. Open the page before you write to a publisher about anything you are calling recent.
Searching by word and by meaning
Both indexes take either.
A keyword search matches the words you typed. Exact, predictable, and blind to any document about your subject that used different vocabulary.
A meaning search compares your query against the stored documents by position, so a relevant page comes back carrying none of your terms. A market discussion that never uses your preferred phrase is what a keyword alert misses and what this finds.
A similarity search reads up to 250 candidate documents and returns 25 rows once grouped. So the result set is a shortlist and never an exhaustive list. And narrowing the query changes which 250 get read at all, which makes a specific query a different search and not a filtered version of a broad one.
What these indexes are not
- Not a social listening feed. Neither index reads social platforms or review sites. They hold published pages, podcasts and video.
- Not complete. No collection of the web is. A page missing from a result is not proof it does not exist.
- Not a permanent archive, in the News Index's case. It reaches one month back beyond the current one.
- Not a journalist database. No reporter profiles, no beats.
When selecting a result for follow-up, check its collection, dates and live source. You can then judge whether it is suitable for the work you are planning now.
See the News Index, the Prospect Index, or how meaning-based search supports industry research.