By Aidan Sowa, CEO and Founder, Frontier Industries | October 2026
A chemical plant is a neighbour before it is a business. It shares a fence, a water table and a wind direction with people who did not choose it, cannot inspect it, and will judge it forever by its worst hour. Everything else about the industry, the margins, the markets, the miraculous usefulness of its products, is commentary on that one relationship. This guide reads the published record on process safety, disclosure rules and the industry's founding catastrophe, and turns it into hiring criteria for an agency. The short version: chemical companies are believed in proportion to what they publish before anyone makes them.
12,000: roughly the number of US facilities filing EPA risk management plans for hazardous processes (US EPA).
1988: the year the industry adopted Responsible Care, its safety and disclosure code, after Bhopal.
15: workers killed at BP's Texas City refinery in 2005, in an explosion investigators traced to culture as much as to equipment (US Chemical Safety Board 2007).
I Discrimen
A compliant plant is not the same as a trusted one
The permit says you may operate. The neighbourhood decides whether operating feels legitimate, and it keeps its own records.
Chemical manufacturers live inside the most complete disclosure regime in industry and still lose public arguments, which tells you the argument is not about data availability. Your emissions are inventoried, your accidents are investigated by a federal board, your risk management plan is public in outline. And yet the fence line community's working theory of your plant is built from odours, sirens and the memory of the last official who said "no cause for concern" about something that later caused concern.
An agency that treats this as an image problem will schedule you a community day and a sustainability film. The right agency treats it as a credibility ledger: every published monitoring result, every prompt incident notice, every honest unknown admitted in a public meeting is a deposit, and the ledger is audited on the worst day of the decade. Deposits cannot be made retroactively. That single fact should organise your entire hiring decision.
Compliance is what the regulator requires. Trust is what the record shows on the days nobody was watching the wording.
II Machina
Your plant's data is already public; the only choice is whether you narrate it
The information machine in chemicals publishes first and invites you to comment.
Consider the machinery arrayed around a single facility. The EPA's Toxics Release Inventory publishes your reported releases annually, in a database journalists mine every year. Roughly 12,000 facilities file risk management plans under the Clean Air Act for their worst case scenarios (US EPA). Refinery fence line benzene monitoring data is posted publicly under federal rules. And if something goes badly wrong, the US Chemical Safety Board will reconstruct your incident in a report and a film, with your internal documents as exhibits, free to every neighbour you have.
In that environment the agency's traditional product, the well timed announcement, is nearly worthless. What has value is narration ahead of the machine: your TRI numbers explained by you before the annual news story; monitoring data published at a cadence the community can set its watch by; incident notices written so plainly the CSB's future summary could quote them without embarrassment.
Survey the industry PR resources built for this programme and the pattern repeats across process industries: the publisher of the primary record sets the terms on which everyone else discusses it. Hire the agency that wants the publisher's chair.
You will be data either way. The choice is between being data with a narrator and data with an accuser.
III Exemplum
Bhopal built Responsible Care; Texas City showed what happens without the culture
The industry's defining catastrophe produced its defining institution, and the sequel showed the institution's limits.
In December 1984, a Union Carbide pesticide plant in Bhopal, India, released methyl isocyanate gas over a sleeping city. Thousands died in the first nights, and the disaster became the permanent reference point for chemical industry trust worldwide. The industry's organised response took four years to arrive and was structural rather than rhetorical: Responsible Care, launched in Canada in the mid 1980s and adopted by the US industry in 1988, committing member companies to safety and environmental codes, community awareness and emergency response planning around every plant. It was, in effect, an admission that the industry's licence would be collective or it would be nothing.
Texas City, twenty years later, demonstrated the failure mode the code exists to prevent. When BP's refinery exploded in March 2005, killing fifteen workers, the Chemical Safety Board found organisational and safety culture deficiencies reaching the top of the company, in a facility whose warnings had been documented internally. The public record and the private record disagreed, and the private one was true.
Give candidates both cases and ask what the agency's role is in each. The sophisticated answer is uncomfortable: in the second case the agency's duty was to tell the chief executive that the gap between the records is a fatal exposure. Hire the firm that has had that conversation.
Codes do not create culture. They create a standard against which a culture can be caught lying, including to itself.
IV Damnum
The cost of a chemical incident is denominated in decades
No industry pays for its worst day over a longer period.
Bhopal's reckoning ran for decades and the site's name remains a synonym forty years on. Texas City produced record safety penalties, a criminal plea, and findings that followed BP into the narrative of its next disaster five years later. In process industries, incidents are not closed when they are settled. They are cited, in permit hearings, in litigation, in insurance renewals and in every subsequent incident's first paragraph, effectively forever.
Price your communications accordingly. The work that prevents that citation chain is not crisis response, it is the decade of boring deposits described above: fence line transparency, community advisory panels with real documents, drill participation the public can watch. Its cost is a modest retainer maintained for years. Its alternative is discoverable in any settlement document: penalties, mandated oversight, and the standing discount applied to everything your company says thereafter.
An agency should be evaluated as a long dated instrument, because that is what it is. The one you hire for the launch will be judged, in a hearing you cannot yet foresee, by the record it built when nothing was happening.
In chemicals, there is no news cycle. There is a record, and it compounds like interest in whichever direction you fund it.
V Genus
The variable is fence line trust, and it is built out of monitoring data
Products differ, processes differ, but the decisive asset is the same at every site: whether the people outside the fence believe the people inside it.
That belief now has an instrumentation layer. Continuous fence line monitoring, community air monitoring networks and published flare and upset reports convert trust from a sentiment into a dataset, and the companies publishing that dataset on their own initiative are quietly rewriting what normal looks like for everyone else. When one plant on a corridor publishes real time readings, the neighbouring plant's silence becomes a statement, and not the one it intended.
An agency for a chemical manufacturer must therefore be, in part, a data publisher: able to render monitoring output, TRI filings and process safety indicators for a lay audience without sanding off the bad readings. Test it with your actual numbers. Hand candidates a quarter of fence line data with one bad week in it and ask for the community summary. The agencies that lead with the bad week understand the assignment. The ones that lead with the annual average understand the old one.
Trust at a fence line is not declared. It is measured, published and argued with, in public, on schedule.
VI Regio
A commodity giant, a specialty maker and an agrochemical firm face different publics
Chemistry is one science and at least four reputational businesses.
| Segment | Audience that decides | Evidence that persuades | Failure that ends you |
|---|---|---|---|
| Commodity and petrochemical producers | Fence line communities, state regulators, buyers | Published monitoring, incident notice speed | A release explained by the parish before the plant |
| Specialty chemical companies | Industrial customers, regulators | Product stewardship files, customer audit results | A customer discovering your incident from the news |
| Agrochemical producers | Farmers, food chain buyers, environmental agencies | Residue and stewardship data, applicator training record | Becoming the named villain of a food scare |
| Consumer chemical brands | Retailers, households, product regulators | Ingredient transparency, recall behaviour | A label scandal in a supermarket aisle |
Each segment's catastrophe has a different narrator. The commodity plant's narrator is a neighbour with a phone. The specialty maker's is a customer's auditor. The agrochemical firm's is a food journalist in another hemisphere. The consumer brand's is a regulator with a press office. Agencies are typically fluent in one narration and dangerous in the others, because the instincts differ: community candour, audit rigour, scientific defence and retail speed are four separate crafts.
The mining companies agency guide in this series covers the sister discipline of living beside extraction. In both industries the closest audience is the sovereign one, which is why the public relations services framework for this programme segments by fence line before product.
Your molecule's market is global. Your molecule's reputation is decided within smelling distance.
VII Occasio
The opening: the transition runs on chemistry, and chemistry is not telling anyone
The energy transition is, at the molecular level, a chemical industry expansion, and the industry is oddly silent about it.
Batteries are chemistry. Hydrogen is chemistry. Carbon capture, insulation, lightweight composites, semiconductor process chemicals, water treatment for every new fab and mine: the buildout's bill of materials reads like the industry's product catalogue. On the American Chemistry Council's accounting, chemistry's total contribution runs to roughly a quarter of US GDP once its downstream role is counted (ACC 2024). Yet public argument about the transition proceeds as if materials arrive by magic, and the industry supplying them remains the default villain of the environmental story.
That is an unclaimed position of real value: the enabler with receipts. The company that publishes its role in verified terms, this many tonnes of insulation, this much battery precursor capacity, these emissions per unit against a stated baseline, gives journalists and policymakers a character the transition story currently lacks, and reframes the fence line conversation at the same time.
Ask agencies to build that case from your actual product slate, with numbers your engineers will sign. If the pitch is adjectives about innovation, they have brought you the industry's current failing in a nicer font.
Every transition technology is a chemical product wearing a costume. The industry that makes them should say so, with data.
VIII Structura
The community advisory panel is the strongest structure in this industry, and the most faked
Chemicals owns a mature trust technology that other sectors are still improvising: the standing community panel, and its condition predicts everything.
A real panel has a decade of minutes: company engineers presenting monitoring data, residents asking questions that occasionally hurt, actions logged and reported back, membership that includes the plant's genuine critics. A fake panel has a bus tour and a barbecue. Journalists, regulators and, after an incident, plaintiffs' lawyers can tell the difference in minutes, because the real one has a paper trail of the company being told things it did not enjoy and acting on some of them.
The agency's structural job is to run the real version and to build its siblings: plant managers briefed like public officials, local spokespeople who are actually local, emergency drill communication rehearsed with the parish or county rather than performed for it. It is civic infrastructure, maintained the way you maintain a flare stack, because the failure mode is identical: needed on the worst day, in working order, with no time to build it.
Evaluate candidates by asking to see a panel's minutes, redacted as needed. The firm that produces minutes with hard questions in them has shown you its best credential. The firm's own team account should let you name who would sit in that room for you.
Trust in this business is minuted or it is marketing. There is no third state.
IX Consilium
Score for fence line fluency, data publishing and the courage to bring bad news
The hiring order, assembled from the cases and the disclosure machinery above.
Fence line fluency first: named communities, named panels, years on the ground at operating sites. Data publishing second: TRI, monitoring and indicator material rendered for lay readers without cosmetics, samples inspectable. Third, and decisive in the interview, evidence of upward honesty: a time the agency delivered news a chief executive did not want, and what happened next.
| Your situation | Agency to hire | Agency to avoid |
|---|---|---|
| Commodity producer with fence line exposure | Community embedded process industry firm | Consumer brand agency |
| Specialty maker selling to audited customers | Stewardship and documentation specialist | Generalist with a sustainability deck |
| Agrochemical producer | Science defence firm with food chain fluency | Agency that treats every study as a messaging problem |
| Any manufacturer with an ageing plant | Crisis drilled firm with CSB literacy | The incumbent agency that has never filed a hard quarter |
Trial finalists on the artefacts that will decide your decade: the community summary of one bad monitoring week, and the first hour script for a release your plant has not had. Then reference check on both sides of a fence line. The second call will tell you more per minute than the first.
Hire the agency your neighbours would describe as straight. In this industry the neighbours are the credit rating that matters.
X Quaesita
Questions chemical company buyers actually ask
What should we publish before an incident forces us to?
Your monitoring summaries on a fixed cadence, your TRI figures with your own explanation attached, your process safety indicators in lay terms, and your emergency notification commitments: who is told, how, within what time. Publishing in the quiet years is what converts an incident statement from a debut into a continuation.
How do we talk to a community after a release?
Lead with what happened, what was measured, what is not yet known and when the next update comes, in that order, and never use the word safe as a conclusion the audience should take from you. East Palestine taught the whole process industry that reassurance delivered ahead of data is read as contempt, however good the intentions. Show readings, name the uncertainty, keep the appointment.
Is Responsible Care membership a communications asset?
Only if you operate it as one: audits summarised publicly, panel minutes kept, performance against the codes reported with failures visible. As a logo it is wallpaper. As a discipline with published evidence it is the industry's one homegrown trust instrument.
How should a plant manager be prepared?
As a public official, because functionally he is one. Training with real documents, authority to release defined classes of information without headquarters approval during an event, and a rehearsal schedule tied to drill season. The first voice a community hears in an incident should be a voice it already knows from the panel meetings.
What does the agency do about our TRI numbers each year?
Owns the calendar and the narrative: data verified with your environmental team, context drafted before the database updates, year on year movement explained including the increases, and local reporters briefed with the full table rather than the favourable rows. The annual TRI story will be written. The only question is whether your explanation is in its first paragraph.
How do we handle activist scrutiny of a specific product?
With the product's file, not with the company's halo: toxicology summaries in plain language, exposure assessments, the regulatory history and the open questions stated. Agencies that answer a product challenge with corporate reputation material concede the only ground that matters.
Which metrics tell us it is working?
Notice speed and accuracy during upsets, panel attendance and question difficulty over time, the share of local coverage that uses your published figures, and grievances closed with published outcomes. Track also the hardest metric: how your plant is described in the first paragraph of someone else's incident story. That paragraph is your true credit rating.
XI Methodus
How this guide was built, and where it is thin
The institutional facts here come from the US Chemical Safety Board's published investigation of the Texas City explosion, the EPA's risk management and Toxics Release Inventory programmes, and the American Chemistry Council's published account of the industry's economic footprint, each cited where used and listed below. The Bhopal and Responsible Care account reflects the standard public record of both.
The caveats are structural. The facility count is an EPA approximation and moves as rules change. Texas City is a refinery case used for a manufacturing audience because the CSB's cultural findings generalise across process industries, but the transfer should be noted. Panel quality, this guide's central structural claim, is real but not publicly measured, so its weight here is argument, not data. The hiring order is labelled inference from the failure record.
XII Incipe
Where Frontier Industries fits in this
Frontier Industries works chemical accounts as a publishing discipline at a fence line: monitoring cadence, panel craft and incident scripts rehearsed before they are needed, built for an industry whose record is public and whose trust is local.
If you are choosing an agency for a chemical company, open with your worst monitoring week and your panel minutes: request an industry campaign consultation or call 401-219-4207.
Sources
1. US Chemical Safety and Hazard Investigation Board, investigations and reports, https://www.csb.gov/ 2. US Environmental Protection Agency, Risk Management Program, https://www.epa.gov/rmp 3. US Environmental Protection Agency, Toxics Release Inventory programme, https://www.epa.gov/toxics-release-inventory-tri-program 4. American Chemistry Council, https://www.americanchemistry.com/ 5. International Council of Chemical Associations, Responsible Care, https://www.icca-chem.org/responsible-care/ 6. US Environmental Protection Agency, East Palestine train derailment response, https://www.epa.gov/east-palestine-oh-train-derailment 7. AIChE Center for Chemical Process Safety, https://www.aiche.org/ccps 8. Society of Chemical Manufacturers and Affiliates, https://www.socma.org/ 9. Institution of Chemical Engineers, process safety resources, https://www.icheme.org/ 10. National Fire Protection Association, research on process and fire safety, https://www.nfpa.org/education-and-research/research