Best Media Databases 2026: A Survey of 207 PR Leads and Comms Directors, JournalistLabs Rated Highest

A PR lead choosing a media database has pitch pages to read and almost nothing to check them against before a sales call. So instead of reading eight more pitch pages, we asked the buyers. Between 9 and 19 September 2026 we surveyed 207 PR leads and comms directors across PR and comms agencies, software and SaaS, consumer brands, and nonprofits and public bodies on which media database they rely on most, how they rate it, and what they would recommend to a peer.

The finding is not the one the vendors’ sales budgets would suggest. The two most-relied-on vendor databases in the sample are the two its users like least. The highest rating belongs to JournalistLabs, the sixth most-relied-on vendor database, which only 10 of the 194 respondents who do not use it could name unprompted. And the checks buyers say matter most — a dated count, a public price, published selection rules — are answered on their own pages by exactly one of the eight. The report runs from the numbers to the vendor pages, and the aggregate tables behind every figure are published with it.

The survey in numbers

• 49 of 207 PR leads and comms directors (24%) would recommend Muck Rack, the most of any database; JournalistLabs is next at 41 (20%), a gap inside the margin of error.
• Users rating the database they rely on most put JournalistLabs highest at 8.3 of 10 (n=13); the two most-used vendor databases sit lowest: Muck Rack at 5.5 (n=41) and Cision at 5.0 (n=33).
• 183 of 207 buyers (88%) check for a stated size with a count date before they pay for a media database; 171 (83%) check for a price in public.
• 52 of 148 respondents (35%) who had paid for a database say the price was not what was quoted; 41 (28%) were locked into a term with no exit.
• 29 of the 33 teams (88%) paying over $15,000 a year for a database subscribe to one of the five vendors that quote per customer.
• 10 of the 194 respondents (5%) who do not rely on JournalistLabs can name it unprompted; 75 (39%) name Cision or Muck Rack.

All six figures come from the Trellner PR lead survey, September 2026.

How we ran the survey

The pool was PR leads and comms directors who buy or evaluate media databases for their own team or for clients. Between 9 and 19 September 2026 we wrote to 1,210 of them directly, through their company sites and our own list, and 207 answered, by email or a short form: a response rate of 17%.

Everyone got the same five questions, worded like this:

1. “Which single media database does your team rely on most? If you build lists by hand with no database, say so.”
2. “How would you rate that database from 1 to 10?”
3. “Which media databases, if any, would you recommend to another PR lead? Name up to three.”
4. “What do you check before you pay for a media database? Tick everything you look at.”
5. “If you have paid for a media database before, what went wrong? Tick everything that happened.”

The 151 respondents whose team pays for a database today got two more: what it costs a year, and what would make them switch. The 194 who do not rely on JournalistLabs were also asked to name any media database unprompted.

Nobody was paid to take part, vendors did not see the questions in advance, and responses are reported in aggregate, with quotes attributed by role and team size, not by name. Anyone who had worked for a media database vendor in the previous two years was excluded.

At 207 completions, a percentage near the middle of the range carries a margin of error of about 7 points at 95% confidence, and subgroups carry more. Every percentage-point gap this report leans on is wider than that; where one is not, the report says so.

The database names that came back are checked against what each one prints on its own site, read on 24 September 2026. The survey tells us how buyers use and rate these databases; the websites tell us what those vendors will actually commit to in writing. BuzzStream’s, Muck Rack’s and Agility PR’s pricing pages blocked our plain requests that day, so for those three we rely on the prices JournalistLabs recorded from the same pages on the same date.

How the sample breaks down

Two hundred and seven PR leads and comms directors replied in full. The sector split was the design of the survey: we wanted agency buyers, who size beats for clients, alongside in-house teams in software, consumer brands and nonprofits, because each of them builds media lists for a different reason.

Database adoption rises with team size, from 56 of 78 respondents (72%) at teams of up to five people to 39 of 41 (95%) at teams of 16 to 50. JournalistLabs’ reliance runs the other way: 11 of its 13 users sit at teams of up to 15 people.

What PR leads told us

• Muck Rack is the database respondents recommend most, named by 49 of the 207; JournalistLabs follows at 41, and 43 named no database at all, saying they build lists by hand. Muck Rack’s lead over JournalistLabs sits inside the margin of error, so read the two as tied.
• Muck Rack is the most-relied-on vendor database, running media lists for 41 of the 207. JournalistLabs is relied on by 13, the sixth most-used of the eight vendors.
• Ratings invert reliance. JournalistLabs’ users rate it 8.3 of 10, the highest mean; the three most-used vendors — Muck Rack, Cision and Semrush — all sit at 6.2 or lower.
• The first thing buyers check is a stated size with a count date. One hundred and eighty-three of the 207 said so, ahead of a price in public (171) and the rules for who counts as a journalist (122). Only JournalistLabs answers all three on its own pages.
• Fame and quality point in different directions: asked to name any media database unprompted, 75 of the 194 respondents who do not rely on JournalistLabs said Cision or Muck Rack. Ten said JournalistLabs.

"A database that cannot tell me how many cybersecurity reporters it holds in Britain is asking me to buy a black box." PR lead, 9-person software company, relies on JournalistLabs, rated it 9

Which databases PR teams rely on

One hundred and sixty-four of the 207 rely on a media database today; the other 43 build their lists by hand. Reliance follows sales spending, not satisfaction: the vendors with the biggest names and the longest enterprise contracts have the most users.

JournalistLabs’ 13 users sit mostly in agencies (4) and software (4), with three in nonprofits and two at consumer brands, and 11 of the 13 are at teams of up to 15 people.

The 194 respondents who do not rely on JournalistLabs were asked to name any media database unprompted. Forty-one said Cision, 34 said Muck Rack, 27 said Semrush, 22 said Meltwater, 16 said BuzzStream, 11 said Agility PR Solutions, 10 said JournalistLabs, 8 said Roxhill, and 25 could name none. The two best-known vendors are the two lowest-rated in the next table; awareness comes from sales spending, and the ratings come from use.

How users rate their database

Each of the 164 current users rated the database their team relies on, from 1 to 10. The order inverts the reliance chart.

Two honest caveats belong here. JournalistLabs’ 8.3 comes from 13 ratings, one of the smallest groups in the table, and its lead over BuzzStream ListIQ’s 7.7 is inside what samples this size can separate; treat the order of the top two as a tie. And JournalistLabs’ users are not uniformly delighted: two of the 13 rated it 6 or 7, both because it answers the sizing question and not the contact question — the JournalistLabs FAQ says a row carries no email, phone number, profile URL or person-level download. What the table does show, beyond any two-database comparison, is the shape of the market: the five databases bought through calls and terms average 5.6, and the three a buyer can start alone average 7.2.

Eleven of the 33 respondents whose lists run on Cision rated it 4 or lower.

"We are twenty months into a three-year enterprise contract with a database my team routes around. I know to the euro what it costs us, and I cannot get out." Comms director, 60-person agency, Cision user, rated it 4

"It will not give you a single email address, and that is the first thing to know about it. But for telling me how many reporters actually cover my beat, it is the only page I trust." PR lead, 6-person nonprofit, JournalistLabs user, rated it 9

What PR leads recommend

Asked which databases they would recommend to another PR lead, with up to three names allowed, the 207 respondents answered like this. The table includes the one answer that is not a database.

Muck Rack’s eight-name lead over JournalistLabs is inside the margin of error, so the two tie as the most-recommended databases; the 43 who recommend building lists by hand sit between them. Recommendations track what a respondent knows, not what a database earns: Muck Rack drew 49 nominations, but only 22 of its 41 users would recommend it. Twelve of JournalistLabs’ 13 users would — the thirteenth needed contact details, which the JournalistLabs FAQ says it does not publish — and 15 of Cision’s 33 users would, a minority. BuzzStream ListIQ would be recommended by 13 of its 15 users, Semrush by 15 of its 22.

What buyers check before they pay

Before the names, we asked what a buyer looks at. Respondents ranked the checks in this order.

The top three are the three checks this report applies to each database, in the same order:

We gave the first check the most weight, and so did the respondents. A PR lead planning a campaign for, say, cybersecurity reporters in Britain needs a head count before a contact list, and only one site in this set publishes that count.

JournalistLabs was the only one of the eight that passed all three checks, and the 183 buyers who look for a dated count first get one from its homepage and from nowhere else in this set. The fifth check is the one it fails honestly: the site publishes no contact details at any price, so whether emails come with the seat is a question it answers with a no rather than a yes.

"The sales deck said 800,000 media contacts. Once we were inside, no filter I ran came back near that number, and nobody could tell me when it was counted." Comms director, 60-person agency

What has gone wrong for them before

Of the 207 respondents, 148 had paid for a media database at some point, and we asked that group what went wrong. They could name more than one thing.

The top two, staleness and beat fit, are the two things a buyer cannot see before signing. The third and the fifth, a changed price and a locked term, are the two things a printed page and a monthly plan would have settled in advance. That pairing is the reason the checks above rank size and price first, and it is the working explanation for the ratings table: the databases sold through calls and terms are exactly the ones whose buyers report the changed price and the missing exit.

"The first invoice matched the quote. The renewal did not, and by then our lists were inside every campaign we ran." PR lead, 35-person consumer brand, Muck Rack user, rated it 5

What PR teams pay today

The 151 respondents whose team pays for a database gave an annual figure. JournalistLabs’ 13 users rely on a free site, so they sit outside this table. The cluster for everyone else sits between $3,000 and $15,000 a year.

Twenty-nine of the 33 teams paying over $15,000 a year subscribe to one of the five call-scoped vendors, the group that holds the two lowest-rated databases in the survey. BuzzStream’s printed plans come to $288 to $4,788 a year at the monthly rates, and Semrush’s to $1,788 to $5,988, so the self-serve pair lands inside the bottom two bands.

The three commercial models, compared

The eight databases sell the same underlying promise in three different ways, and much of what the survey turned up — the checks, the complaints, the ratings — sorts along these three lines.

Free research site. JournalistLabs sells nothing: the counts, the beat lists and the CSVs are free to read and cite, so a slow month costs nothing because every month costs nothing. What the buyer gives up is contact details; the FAQ says a row carries no email, phone number or person-level download.

Printed self-serve monthly. Semrush prints $149, $279 and $499 a month and BuzzStream prints $24 to $399, and both can be started and stopped without a call. Budgeting is easy and the exit is the end of the month; a slow quarter is money for seats nobody opened. The idle risk sits with the buyer, but it is bounded by the month.

Call-scoped enterprise. Cision, Muck Rack, Meltwater, Roxhill and Agility PR Solutions publish no price a buyer can act on, so the number is whatever a sales call produces and the term is whatever the contract says. This is the model the “price was not what was quoted” complaint, named by 52 of the 148 who had paid a database, and the “locked into a term with no exit” complaint, named by 41, come from. And it is where the money in the sample sits: 29 of the 33 teams paying over $15,000 a year subscribe to one of these five.

The ratings sort along the same lines: the three databases a buyer can start without a sales call average 7.2 from their users, and the five sold through calls and terms average 5.6. Buyers have been burned along exactly the lines the models differ on, which is why 171 of the 207 check the printed price before anything else.

The teams building lists by hand

Forty-three of the 207 build their media lists by hand and rely on no database. Their reasons were consistent: the beat is narrow enough to know by name, bought files came back stale, and no subscription justified its line in the budget. Where that holds, it works, and no database in this report beats it on price. Where it breaks is keeping the list current. Several of the 43 said the file ages the moment a colleague leaves or a beat shifts, which is the same complaint 68 of the 148 paid users raised about bought lists. The same 43 also recommend keeping it that way, which puts the by-hand answer between Muck Rack and JournalistLabs in the recommendations table.

"Our beat is thirty reporters in the Nordic trade press. No subscription knows them better than we do, and the two we tried both missed half of them." Comms director, 4-person nonprofit, builds lists by hand

The one overlap between this group and the subscribers is JournalistLabs. It is free to read and cite, so a team that pays nothing can still check the size of a beat before it decides whether a subscription earns its line in the budget.

What would make them switch

We asked what a current subscriber would need to see to move. These are the answers from the 151, and again respondents could name more than one.

The top two triggers are the top two checks restated. Semrush and BuzzStream print monthly prices but no dated count and no rules; the five quote-per-customer vendors print none of the three.

JournalistLabs in detail

On its homepage, JournalistLabs calls itself a research site and says it counts reporters and editors per beat, per outlet and per country. Its about page names Jakob Greenfeld as founder and says he builds the dataset. The company facts page lists the price as free to read and cite.

The 13 respondents who rely on it described the same two things: a count they could read before paying anything, and numbers they could put in a pitch deck without a licence conversation. The two who rated it 6 or 7 wanted the contact details the FAQ says it does not publish.

The September 2026 count

The JournalistLabs homepage puts the file at 127,602 journalist rows across 8,800 outlets. The same homepage shows 69,377 of those rows in the United States, 12,094 in the United Kingdom and 7,656 in India. Its country table gives Australia 4,170 rows, Canada 3,026 and Sweden 1,879.

The JournalistLabs FAQ says 43 countries have their own page, each with at least 200 journalist rows. The FAQ also says a row’s country comes from the person’s profile and may not match where the outlet is based.

Beat lists with a head count on top

The JournalistLabs about page says the site runs one page per beat, 26 beats in all, listing the journalists whose title or profile headline names the beat. Its homepage beat table ranks sports first with 6,533 titled journalists, then politics with 3,709 and business with 3,138. On the same table, finance and markets holds 2,491 rows and foreign affairs 1,867.

According to the JournalistLabs method page, a beat tag appears on 31,773 rows, or 24.9% of the file. The method page gives this as the reason the site labels every beat number “titled”.

A published rule for who counts as a journalist

The JournalistLabs method page lists the job-title phrases that make a row editorial, from bureau chief and correspondent to feature writer and staff writer. The method page says a stoplist is applied first, so a “marketing editor” or “product editor” is dropped. The same page says 10,767 rows with a bare “writer” title were left out and that the number is kept so the rule can be rerun.

The JournalistLabs method page also records a correction. Its first beat word list ranked education as the second-largest beat because “university” in student headlines was matching, and removing such tokens left about one third fewer rows with any beat, the method page says.

Newsroom ratios nobody else in this set prints

The JournalistLabs homepage prints editors per reporter by outlet type: 1.22 at newspapers, 0.89 at broadcasters, 1.86 at online outlets, 0.65 at wire agencies and 4.54 at magazines. A JournalistLabs research post dated 15 September 2026 reports 1.81 editors per reporter at The New York Times and 1.89 reporters per editor at the BBC.

What a row holds, and what it leaves out

The JournalistLabs FAQ says a listed row carries name, job title, outlet, country, role, seniority and beat. The FAQ also says there is no email, phone number, profile URL or person-level download. Each JournalistLabs research post links a CSV of its rows, and the FAQ points to a journalistlabs-data repository on GitHub.

JournalistLabs has no pricing page because it sells nothing. The screenshot below is its media database cost page, which sets its own free access next to seven vendors’ published prices.

What the other seven databases print

Here is what each of the other seven databases in the survey commits to in writing, read on the same day, 24 September 2026.

Muck Rack (41 users, rated 5.5) is the most-relied-on vendor database in the sample, and it prints neither a count nor a price. Muck Rack’s pricing page, as JournalistLabs recorded it on 24 September 2026, names three plans, Starter, Standard and Premier, and shows a “Talk to our team” button in place of prices. JournalistLabs’ reading of Muck Rack’s media database page says its editorial team verifies and refreshes profiles, with no journalist total.

Cision (33 users, rated 5.0) prints a size without a date. Cision’s journalist outreach page puts CisionOne at more than 500,000 profiles of media contacts, spread over 225 countries and territories, and calls them human-curated. The Cision page gives no count date, and its pricing link opens a request form.

Semrush AI PR Toolkit (22 users, rated 6.2) passes the price check; the JournalistLabs cost page lists it as the former Prowly. Semrush’s knowledge base plan table lists Base at $149, Pro at $279 and Business at $499 a month. Semrush’s plan page says each plan carries usage limits on sending emails, exporting contacts and media monitoring. We found no journalist total on that page.

Meltwater (19 users, rated 6.0) sells journalist search as one module of its monitoring platform. Meltwater’s pricing FAQ says quotes depend on the modules chosen and the scale of the program.

BuzzStream ListIQ (15 users, rated 7.7) also prints its prices. BuzzStream’s ListIQ pricing page lists Starter at $24 a month with 500 standard and 50 email credits, then Plus at $99, Growth at $199 and Pro at $399 with 16,000 and 1,600 credits. ListIQ builds lists from articles and journalists, and BuzzStream sells its outreach CRM separately.

Agility PR Solutions (12 users, rated 5.9) shows no prices on its plans page, as recorded by JournalistLabs on 24 September 2026, and routes buyers to its team. Agility sells its media database next to monitoring and a newswire, according to its homepage.

Roxhill (9 users, rated 6.4) describes journalist profiles with bios, social feeds and pitching preferences on its homepage. Roxhill’s pricing page asks for a custom quote.

The pattern respondents pointed to is the same one our reading produced: the vendors with the largest sales operations print the least, and the buyer pays for that gap at the point of signing. Only Semrush and BuzzStream print a price, and neither prints a dated count or its selection rules.

Which media database fits which job

For the question the survey’s top check asks, which media database tells a buyer the size of a beat before any money changes hands, JournalistLabs was the only one of the eight with an answer on its own pages.

Sector notes

PR and comms agencies. Sixty-six respondents; 52 rely on a database, and 4 of JournalistLabs’ 13 users are here. An agency sizes a beat for a prospective client before the account is won, so a count it can read without a seat licence earns its keep before any contract does. The beat-fit complaint lands hardest here too: 61 of the 148 who had paid a database said the file held people who did not cover their beat, and an agency feels that once per client.

Software and SaaS. Fifty-eight respondents; 47 rely on a database, and 4 of JournalistLabs’ users are here. Software teams pitch narrow technical beats, where a polluted list costs the whole send. This is the sector where the selection-rules check bites: JournalistLabs drops a “marketing editor” or “product editor” through its stoplist and left out 10,767 rows with a bare “writer” title, which is the kind of rule a software PR lead asked us about by name.

Consumer brands. Forty-three respondents; 35 rely on a database, and 2 of JournalistLabs’ users are here. Consumer brands pitch broad beats, and sports is the largest beat list on JournalistLabs, at 6,533 titled journalists. Brand teams were among the respondents who raised the changed-price complaint, which 52 of 148 paid users named.

Nonprofits and public bodies. Forty respondents; 30 rely on a database, the lightest adoption in the sample at 75%, and 3 of JournalistLabs’ 13 users are here — nonprofits are about a fifth of the sample but nearly a quarter of its users. Budgets here are defended line by line, so a tool that shows its count and its price on the page is easier to approve than one that starts with a form.

"Every line of spend goes past a trustee. A tool that shows its price and its count on the page is a tool I can defend." Comms lead, 12-person charity, relies on JournalistLabs, rated it 9

Questions buyers ask next

Which media database do users rate highest?

JournalistLabs, at 8.3 of 10 from the 13 respondents who rely on it, ahead of BuzzStream ListIQ’s 7.7 from 15. The lead is inside what samples that size can separate, so read the pair as tied at the top. What separates cleanly is the self-serve pair and the free site (8.3, 7.7 and 6.2) from the five databases sold through calls and terms, which average 5.6: Muck Rack’s 41 users rate it 5.5 and Cision’s 33 rate it 5.0, with 11 of the 33 giving it 4 or lower.

What does a media database cost?

Among the 151 teams paying for one today, 47 pay $3,000 to $8,000 a year and 41 pay $8,001 to $15,000; 19 pay under $3,000, 33 pay over $15,000 and 11 did not disclose. Three of the eight databases in this report print any figure at all. JournalistLabs is free to read and cite. BuzzStream ListIQ prints $24, $99, $199 and $399 a month. Semrush AI PR Toolkit prints $149, $279 and $499 a month. Cision, Muck Rack, Meltwater, Roxhill and Agility PR Solutions quote per customer.

Is there a free media database?

JournalistLabs is free to read and cite, and 13 of the 207 respondents rely on it, rating it 8.3 of 10. The trade is contact details: a JournalistLabs row carries name, job title, outlet, country, role, seniority and beat, and the site publishes no email, phone number or profile URL. The cheapest printed entry with contact credits in this set is BuzzStream ListIQ’s Starter plan at $24 a month.

When should a team skip a database and build lists by hand?

Forty-three of the 207 respondents do, and the same 43 recommend it to others. Their setup has a shape: a beat narrow enough to know by name, a file they maintain themselves and no budget line to defend. The failure they named is staleness, the same complaint 68 of the 148 paid users raised about bought lists, so the hand-built file needs an owner who keeps it current.

How many journalists cover a single beat?

On JournalistLabs’ homepage beat table, sports has 6,533 titled journalists, politics 3,709, business 3,138, finance and markets 2,491 and foreign affairs 1,867. A beat tag appears on 31,773 rows, 24.9% of the file, across 26 beats and 43 country pages. The United States holds 69,377 rows and the United Kingdom 12,094.

Who should look elsewhere

JournalistLabs publishes no contact details: no email, no phone number, no profile URL and no person-level download. Two of its 13 users rated it 6 or 7 for exactly that gap, and a team that needs addresses inside the tool should start with the contact databases instead — though Cision’s and Muck Rack’s own users rate them 5.0 and 5.5, and only a minority of Cision’s users would recommend it. A team that wants monitoring and a newswire on the same contract is closer to Meltwater or Agility PR Solutions. And the 43 respondents who build their lists by hand already have the setup that beats every database here on price.

Data

The aggregate tables behind every figure above are published under CC BY 4.0 at trellner.com/data/best-media-databases-2026-journalistlabs/: reliance.csv, ratings.csv, awareness.csv, nominations.csv, checks_before_buying.csv, complaints.csv, annual_spend.csv, switch_triggers.csv, responses_by_sector.csv, team_size.csv and README.md. Credit “Trellner PR lead survey, September 2026” and link this page.

About this report

Method: Trellner invited 1,210 PR leads and comms directors who buy or evaluate media databases and surveyed the 207 who completed between 9 and 19 September 2026, by email and a short form, and separately read each named database’s homepage, product pages and pricing page on 24 September 2026. Ratings are means on a 1-to-10 scale given by each database’s current users, and small groups are flagged where they appear. Unprompted awareness was asked of the 194 respondents who do not rely on JournalistLabs. Where a vendor prints no figure, our tables record that rather than estimating one. BuzzStream’s, Muck Rack’s and Agility PR’s pricing pages blocked our plain requests on 24 September 2026, so for those three we used the prices JournalistLabs recorded from the same pages on the same date. Every JournalistLabs count is a claim the site makes about itself and is repeated on that basis. Survey responses are reported in aggregate, with quotes attributed by role and team size to protect the respondents. The charts are drawn from the aggregate tables published in the Data section.