A media monitoring checklist is a list of setup decisions you make before switching on any tool: which names and terms to track, which sources matter, who gets alerted and when, how sentiment is judged, and what the monthly report shows. Getting these right on day one prevents the two usual failures, which are drowning in irrelevant mentions or missing the one story that mattered.
What should you track besides the brand name?
The brand name alone is too narrow and, if it is a common word, too noisy. Build a keyword map in groups. Each group serves a different purpose and often a different person in the team.
- Brand: company name, common misspellings, product names, website domain, social handles
- People: founders, spokespeople and executives who speak in public
- Competitors: their brand names and main products
- Industry: category terms and regulatory topics that affect the business
- Campaigns: hashtags and slogans from current or upcoming campaigns
- Risk: product name plus words like recall, lawsuit, outage, scam or complaint
How do you build Boolean queries that stay clean?
Most monitoring tools, including Meltwater, Brandwatch, Talkwalker, Mention and Brand24, accept Boolean logic. A brand called Nova, for example, needs exclusions, or the feed will fill with astronomy and car models. A query might read: (“Nova Pay” OR “NovaPay” OR novapay.com) AND NOT (supernova OR “Chevy Nova”).
Test every query against the last 30 days of data before you save it. Read the first 50 results. If more than a few are irrelevant, add exclusions or require a context word such as payment, app or fintech within a few words of the brand name, using a NEAR operator if the tool supports it.
Language and local spelling
For brands working across several countries, add translated product terms and local spellings. Turkish, German and Russian coverage, for instance, often uses different forms of company names or adds suffixes. Many tools allow a language filter per query, which keeps reporting by market simpler.
Which sources belong in scope?
Decide this with the client rather than ticking every box. News sites and trade publications matter for most brands. Social platforms matter differently depending on audience. Forums, review sites and podcasts are often overlooked and are where some of the most honest opinions sit.
- Online news and trade media, plus print and broadcast if the budget covers it
- X, LinkedIn, Instagram, TikTok, YouTube and Facebook public pages
- Reddit and niche forums relevant to the category
- Review platforms such as Trustpilot, Google reviews, G2 or app stores
- Podcasts and newsletters, often through transcript search
Note what the tool cannot cover. Private groups, closed communities and most of Facebook’s personal content are outside every commercial tool. Write that in the setup document so nobody assumes total coverage.
Who gets alerted, and how fast?
Alert design is where most setups fail. Sending every mention to everyone creates noise and people mute the channel within a week. Build tiers instead.
A crisis tier sends an immediate notification, through email, Slack or Microsoft Teams, when risk keywords appear with the brand, or when mention volume rises sharply above the daily average. A daily digest goes to the PR lead with new articles and high-reach posts. A weekly summary goes to the wider marketing team.
Defining a spike
Use the tool’s anomaly or spike detection where it exists. If it does not, set a rule yourself: for example, alert when hourly mentions reach three times the trailing 7-day hourly average. Adjust after the first month based on how many alerts were useful.
How should sentiment be scored?
Automated sentiment works reasonably on plain statements and poorly on sarcasm, mixed opinions and many non-English texts. Treat the tool’s score as a starting point. Review a sample each week, around 50 mentions, and correct the labels manually. Some tools learn from these corrections.
Agree on definitions with the client. Is a neutral news article about a funding round positive? Is a customer complaint that was resolved publicly negative? Written definitions keep monthly comparisons consistent even when a different person does the review.
What goes into the first report?
The first month is a baseline, not a verdict. Show total mention volume by source, share of voice against two or three competitors, sentiment split with sample quotes, top outlets and authors, and any notable stories. Keep it to a few pages.
- Volume trend by week and by source type
- Share of voice compared with named competitors
- Sentiment breakdown with five representative quotes
- Top ten articles or posts by estimated reach
- Issues flagged and the action taken
Avoid advertising value equivalency. It multiplies coverage size by ad rates and has been criticised for years by measurement bodies such as AMEC. Report outcomes like message pull-through, key outlet coverage and referral traffic instead.
Ongoing maintenance most teams skip
Queries age. New products launch, executives change and competitors rebrand. Schedule a quarterly query review: add new terms, retire old campaign hashtags and recheck exclusions. Also export your data regularly, because some tools keep only a limited history depending on the plan.
How do you hand the setup over to the client?
Monitoring only helps if the people receiving alerts know what to do with them. Write a one-page response guide alongside the setup document. It should say who decides whether a story needs a reply, who drafts the statement, who approves it, and how quickly each step should happen for different alert tiers.
Run a short walkthrough with the client team in the first week. Show them the dashboard, explain the queries and exclusions, and agree on where they can request changes. Teams that understand how the feed is built trust it more and use it more often.
- Response guide with roles, approval steps and target response times
- Shared document listing every query, exclusion and source setting
- A named contact for query changes and alert tuning
- Agreed format and date for the monthly report
Common setup errors to avoid
The most frequent errors are easy to prevent: alerts sent to a personal inbox of someone who later leaves, queries saved without testing, competitor tracking added but never reported, and sentiment scores shown without any manual review. Each of these turns a useful system into noise within a few months.
Measure the monitoring itself
Once a quarter, count how many alerts led to action. If most did not, tighten the queries.
Where to start this week
Write the keyword map on one page in the six groups above, then build and test the brand query against 30 days of history. Before you set any alert, decide with the client who should receive crisis notifications and on which channel. Those three steps take a day and set the foundation for everything the tool will report afterwards.
Keep reading: Media Monitoring · Digital PR