Why competitor news monitoring fails on keywords alone
Competitor news monitoring is not searching a rival’s name every morning. Company launches, defense awards, and earnings hit the wire in the same week as political rivals fight for headline share. Keyword q alone cannot keep those lanes clean.
Search Lockheed, RTX, or defense and you also match namesakes, sports metaphors, and consumer brands. Search Trump or Biden and you also match family members, entertainment rows, and caption noise. The rivals you mean stay hard to compare side by side.
Try it live: Articles sandbox · Docs: Entities · Plans: API pricing
This guide answers:
What is competitor news monitoring on a news API?
How do you build a rival company watchlist with entities?
How do you compare political rivals like Trump and Biden?
How do you compare coverage volume across competitors?
How do you keep earned coverage vs press-release noise?
Which recipes fit company CI vs political CI vs Stories?
Param encyclopedias live in the Company News API guide and Entities docs. Product briefings belong on Monitor Defense Industry News or Monitor US Politics News. This page owns the API competitive-intelligence workflow - companies and people as rivals.
What competitor news monitoring means on a news API
On Perigon, competitors are whoever you need to compare in coverage - primes, brands, or political figures. The workflow is the same shape:
1) Pick a stable rival set.
2) Resolve each rival to a durable ID (companyId / companySymbol, or personWikidataId).
3) Run the same Articles (/v1/all) recipe once per rival in one date window.
4) Compare volume and themes; optionally open Stories when one move reprints everywhere.
You filter who was linked on the article. You do not scrape competitor websites. You do not call another vendor’s news API.
| Rival type | Resolve with | Filter on Articles / Stories |
|---|---|---|
| Company / brand | Companies API | companySymbol, companyId, companyDomain |
| Political figure | People API | personWikidataId (prefer) or personName |
| Event language | - | Optional q (award, earnings, debate) |
| Earned-only | - | excludeLabel |
| Reprint collapse | - | Stories with the same entity fence |
How to build a rival company watchlist with entities
Resolve each rival once on the Companies API (/v1/companies/all). Prefer the parent listed company and a primary companySymbol. Skip subsidiaries and namesakes unless that unit is the desk’s target.
Example A - Defense primes as rivals
Live Companies lookup (not a Signals tutorial):
| Rival (parent) | Prefer | companyId | Symbol |
|---|---|---|---|
| Lockheed Martin Corporation | Parent - not Rémy Martin / Martin Professional | 2fa32c247a6f4d83be3c163cc685691b | LMT |
| RTX (US aerospace / defense) | companySymbol RTX - name “RTX” can return Danish RTX A/S | (use symbol; skip f531cda3de2c47128cee041eb5327d83) | RTX |
| Northrop Grumman Corporation | Parent - not Sperry Marine alone | 7ba58a4019cf48afb56c710b0e8d0ef2 | NOC |
| The Boeing Company | Parent - not Boeing Store / BECU | c825868b46614e598257013cf9f39c03 | BA |
| General Dynamics Corporation | Parent - not Mission / Land Systems / Electric Boat alone | a61d0b7101cf411ebcfbacd479586be7 | GD |
Apply it
Store id + symbol in config.
Pass companySymbol set to LMT (or Lockheed’s companyId) on Articles.
Repeat for RTX, NOC, BA, and GD.
Add q only for move type (award, earnings, partnership).
Example recipes (defense)
| CI question | Recipe |
|---|---|
| All Lockheed-linked news | companySymbol=LMT, sort by date |
| Lockheed awards only | companySymbol=LMT + q=award OR contract |
| RTX vs LMT volume this month | Two queries, same from/to, compare numResults |
| Boeing commercial + defense noise cut | companySymbol=BA + excludeLabel Non-news |
| GD shipbuilding desk | companyId for General Dynamics Corporation (avoid Electric Boat-only unless needed) |
| Five-prime morning digest | Five parallel queries → Summarizer per rival |
Name collisions are the failure mode. “Martin” or “RTX” without entities are research strings, not keys. Refresh IDs when tickers or structure change.
Example B - Brand and sponsor rivals (pattern)
The same pattern works outside defense. Resolve each brand on Companies, then run parallel companySymbol or companyId watches - for example Nike (NKE) vs a sportsbook or apparel rival on your book. Keep one query per brand in the same window. Domain filters (companyDomain) help when the rival is private or the ticker is ambiguous. Full ticker/domain cookbook stays in the Company News guide.
How to compare political rivals like Trump and Biden
Political competitors are people, not tickers. Resolve each on the People API (/v1/people/all), store personWikidataId, and run parallel Articles queries. Deeper person methodology lives in the political rival mentions guide - the worked pair below is the CI example for this page.
Example C - Trump vs Biden (live People lookup)
| Rival | Prefer this row | personWikidataId |
|---|---|---|
| Donald Trump | US politician / President row - not Donald Trump Jr. | Q22686 |
| Joe Biden | 46th President row - not Hunter Biden, Ashley Biden, or singer “Joe” | Q6279 |
People search for “Donald Trump” also returns Melania, Ivana, and Ivanka. “Joe Biden” also returns Hunter, Ashley, Valerie Biden Owens, and unrelated Joe rows. Pin Wikidata IDs or the wrong human enters the feed.
Apply the pair
Feed A: personWikidataId set to Q22686.
Feed B: personWikidataId set to Q6279.
Same from / to window for both.
Optional fence: category set to Politics and sourceCountry set to us.
Optional q: debate, executive order, campaign - event language only.
Political volume (30 days)
pubDate 2026-06-23 through 2026-07-23:
| Filter | About how many articles |
|---|---|
| Trump Q22686 | ~980,000 |
| Biden Q6279 | ~75,000 |
| Trump + Politics + US sources | ~290,000 |
| Biden + Politics + US sources | ~38,000 |
Round large totals; very large person feeds can sit near caps - prefer relative wording. In this window Trump-linked volume sat roughly an order of magnitude above Biden overall, and still several times higher inside the US Politics fence. Use that as triage for coverage heat - not as an election scoreboard.
Political recipe table
| CI question | Recipe |
|---|---|
| Open Trump mention firehose | personWikidataId=Q22686 |
| Open Biden mention firehose | personWikidataId=Q6279 |
| Fairer US politics desk compare | Each ID + category=Politics + sourceCountry=us |
| Debate-week slice | Each ID + q=debate + same from/to |
| Shared narrative heat | Stories + personWikidataId (one ID per poll) |
| Cut opinion roundups | excludeLabel Opinion or Non-news on either feed |
Coverage shape (historical overviews - domains only): news.yahoo.com-style politics commentary linked to Trump; nwaonline.com-style White House visit coverage linked to Biden. Treat as pattern checks. A recent Stories shape for Q22686 can look like a multi-outlet foreign-policy or defense cluster that also mentions Trump - useful when the rivalry is a shared narrative, not a single URL.
For continuous politics product briefings, use Monitor US Politics News. For who wrote the coverage, use election journalist guides.
How to compare coverage across competitors
Whether rivals are companies or people, the rule is the same: one query per rival, identical window. Do not OR five tickers or two surnames into one bag if you need share-of-voice numbers.
Company Articles volume (same 30 days)
| Rival symbol | About how many articles |
|---|---|
| LMT | ~19,000 |
| NOC | ~9,200 |
| RTX | ~5,400 |
Lockheed-linked coverage outpaced Northrop and RTX - triage, not proof of wins.
Company Stories compression (same window)
| Rival | Stories clusters (~) |
|---|---|
| LMT | 61 |
| NOC | 26 |
| RTX | 19 |
Tens of thousands of articles become tens of narratives. Use Articles for volume; use Stories when an earnings print or award reprints (RTX outlook raise; Northrop Hawkeye award). Adjacent science clusters can appear on a prime symbol - tighten with q. Full move-detection sits in the competitor moves / Stories guide.
Side-by-side: company vs political competitors
| Defense primes | Trump vs Biden | |
|---|---|---|
| Entity | companySymbol / companyId | personWikidataId |
| Typical 30-day scale | Thousands–tens of thousands of articles | Tens of thousands–hundreds of thousands |
| Stories role | Earnings / contract moves | Shared policy or campaign narratives |
| Common noise | Namesake firms, subsidiaries | Family members, entertainment Joes |
| Desk fence | Optional sourceGroup / q | Politics + US sourceCountry |
What company coverage often looks like
Entity filters typically return contract and program wires. Historical pattern checks (domains only): marketwatch.com / theatlasnews.co style F-35 award framing for Lockheed; militaryleak.com style F135 engine funding for RTX.
For themes, skim titles per rival feed - or pass each set into Search Summarizer. Summarizer theory stays on that guide.
How to keep earned coverage vs press-release noise
Rival IR and campaign communications publish releases that index like news. For third-party coverage, add excludeLabel (Press Release, Paid Content, or Non-news) on the same entity query. Keep a second feed for official releases if you still need them.
| Goal | Mechanism |
|---|---|
| Company allowlist | companySymbol or companyId |
| Person allowlist | personWikidataId |
| Event language | Optional q |
| Earned-only | excludeLabel |
| Publisher fence | Optional sourceGroup |
| Incremental poll | from / to or addDateFrom |
Mixed stack example:
companySymbol=LMT + q=contract + excludeLabel Press Release - awards discussed by third parties, not only Lockheed’s wire.
Political stack example:
personWikidataId=Q22686 + category=Politics + sourceCountry=us + excludeLabel Non-news - US politics desk, less caption junk.
End-to-end CI playbooks
Playbook 1 - Five-prime defense morning brief
Resolve LMT, RTX, NOC, BA, GD (table above).
Poll Articles per symbol for the last 24 hours (addDateFrom or from/to).
Rank rivals by numResults; open the top two feeds.
For any spike, switch to Stories with that symbol and read articleCount.
Optional: Summarizer on the spike set for a six-bullet IR note.
Playbook 2 - Trump vs Biden weekly share-of-voice
Confirm Q22686 and Q6279.
Run both person filters for the last 7 days with Politics + US sources.
Record relative volume (do not treat as polls).
Diff top topics or skim titles for theme gaps.
If one shared narrative dominates, follow it on Stories with one ID at a time.
Playbook 3 - Award week for two primes
Parallel: companySymbol=LMT and companySymbol=NOC.
Add q=award OR contract OR Navy on both.
Compare counts; open Stories on the louder symbol.
excludeLabel Press Release if you want earned desks only.
When this workflow vs other Perigon surfaces
| Situation | Prefer |
|---|---|
| Rival-company and rival-person API compares | This page |
| Deep person-only methodology | Political rival mentions guide |
| Earnings / contract as one narrative | Competitor moves / Stories guide |
| Full company / ticker / executive params | Company News guide + Entities docs |
| Defense or US politics product UI | Monitor Defense / Monitor US Politics |
| Who wrote the politics coverage | Election journalists guides |
| Editorial calendar gaps | Content strategy news API guide |
| News API vendor prices | Pricing comparison (separate SERP) |
Beat scale at a glance
Company rivals and political rivals both need entity IDs - keywords fail differently (namesakes vs family).
In 2026-06-23–2026-07-23, LMT / NOC / RTX article volumes differed by multiples; Stories stayed in the tens of clusters.
In the same window, Trump vs Biden person volume differed by roughly an order of magnitude; Politics + US sources cut both feeds while preserving the gap.
Contract wires and candidate mention firehoses are evergreen shapes; re-run stats when you ship.
Key takeaways
Competitor news monitoring = resolve rivals + parallel Articles filters - companies or people.
Defense example: LMT, RTX, NOC, BA, GD - skip namesakes and Danish RTX A/S.
Political example: Trump Q22686 vs Biden Q6279 - skip Jr., Hunter, and other near matches.
Compare with one query per rival; use Stories when reprints are the problem.
Stack excludeLabel for earned media; use Signals Monitor pages for product briefings.
Get started: Articles sandbox · Entities · API reference · API pricing
Top comments (0)