Headline counts
Four numbers frame the market
Every figure is a dated public capture from filings, investor materials, or industry research. Analytical constructs such as map positions and the query ladder keep the est. label throughout this edition.
Executive read
Four patterns explain how this market behaves
The manufacturers are the evidence base, not the audience. The useful unit of research is the pattern a marketer inside a filtration business can compare with their own funnel, lookup tooling, and channel data.
The machine is sold once, the filter is sold for its life
Replacement filtration is a razor-and-blade market where the blade sells for 10 to 15 years after the razor. The aftermarket carries the recurring, higher-margin profit pool, which is why every position on the map comes down to how a player defends the reorder.
The cross-reference tool is the marketing
A filter buyer does not shop a brand; they hold a spent part and look up its number. The tool that resolves a competitor part number into an equivalent owns the highest-intent moment in the category, doubling as lead capture, programmatic SEO surface, and defensive moat.
Atmus has to build the demand it used to inherit
Cummins listed Atmus in May 2023 and completed the split-off in 2024. A roughly 80% aftermarket pure-play now funds its own brand search, content, and attribution for the first time, making it the live strategic event every competitor is reacting to.
The reorder trigger is becoming a data fight
Condition-based replacement is displacing fixed intervals. OEMs stream machine data through telematics platforms; independents answer by putting the sensor on the filter itself. Whoever owns the signal that a filter is spent owns the timing of the sale.
The field
One reorder stream, six ways to defend it
Around the reorder sit two scaled independents, a private German group, a diversified giant, the OEM genuine-parts programs, and the will-fit importers. Each relates to the replacement stream differently, and the marketing follows the relationship.
The most digitally mature independent: an open cross-reference, Shop e-commerce, and Filter Minder sensors, on $3.5B FY2024 net sales with aftermarket parts as the profit engine (reported).
Fleetguard is designed into Cummins engines and inherits the genuine reorder; since the 2024 split-off the company has had to stand up its own demand engine.
The family-owned German group, roughly EUR 5B in sales (reported), bought WIX Filters and Purolator in 2016 and runs one of the best open cross-reference tools in the category.
Parker Hannifin acquired Clarcor for about $4.3B in 2017, taking Baldwin Filters with it: an aftermarket heavy-duty brand competing on cross-reference, availability, and price below OE.
Caterpillar, John Deere, and Volvo sell genuine filters through captive dealer portals. High friction, high lock-in, defended with warranty leverage and value sub-brands rather than open-web search.
Low-cost copies sold through marketplaces on price alone. They free-ride the incumbents' part numbers, publish no cross-reference of their own, and absorb the most tariff exposure.
Market map
The interesting tension sits in the top band
Two forces decide who wins: how dependent a player is on the reorder stream, and how digitally it captures the reorder. The top band holds players who all live on replacement filters yet differ wildly in how they defend the part-number search.
Where living on the reorder meets the digital path that captures it
Positions are analyst estimates (est.) synthesized from public revenue mix, product pages, and each player's digital footprint. They are not survey data or performance scores.
- Atmus / Fleetguard: digital maturity 46 out of 100; aftermarket dependency 83 out of 100.
- WIX: digital maturity 77 out of 100; aftermarket dependency 79 out of 100.
- Donaldson: digital maturity 71 out of 100; aftermarket dependency 65 out of 100.
- MANN+HUMMEL: digital maturity 57 out of 100; aftermarket dependency 71 out of 100.
- Baldwin: digital maturity 37 out of 100; aftermarket dependency 76 out of 100.
- Cat parts: digital maturity 65 out of 100; aftermarket dependency 53 out of 100.
- John Deere: digital maturity 55 out of 100; aftermarket dependency 48 out of 100.
- Volvo: digital maturity 43 out of 100; aftermarket dependency 46 out of 100.
- Will-fit imports: digital maturity 18 out of 100; aftermarket dependency 85 out of 100.
- Parker: digital maturity 29 out of 100; aftermarket dependency 36 out of 100.
The lookup
The funnel is inverted and the tool owns it
The buyer holds a spent filter with a number printed on it. They are not researching; they are matching. Value rises as the query gets more specific, and the cross-reference tool that resolves a rival's number wins the sale at the moment of intent.
- Donaldson Open cross-reference + e-commerce Broadest catalog of the independents; lookup feeds Shop.Donaldson directly
- Atmus / Fleetguard Open cross-reference + FIT monitoring Heavy-duty depth tied to the Cummins-designed installed base
- WIX (MANN+HUMMEL) Open cross-reference Among the best lookup UX in the category; strong North American parts-store coverage
- Baldwin (Parker) Open cross-reference catalog Will-fit positioning, distributor-oriented
- OEM programs Dealer-gated parts portals Genuine-only portals like Parts.cat.com; high friction, high lock-in
- Will-fit imports None of their own Copies get discovered because they map to an incumbent's part number
The query ladder, by commercial value
Illustrative intent index (est.): an analytical ranking of how close each query type sits to a purchase, not measured search volume.
The part number as the unit of buying. The most specific query is still the most valuable one, and the long tail of part-number pages remains the category's core search asset.
Who owns the trigger. Fixed service intervals give way to condition-based replacement, and the filter-spent signal now sits in OEM telematics or on the independents' own filter sensors.
AI answer engines are the new leak: if a model resolves a cross-reference in the chat, the click never reaches the brand's tool. Feeding the answer with structured part data becomes the search defense of 2026.
Demand
Five forces set demand through 2026
Filter demand is a function of how many machines exist, how hard they run, and what spec they require, plus who decides when the filter gets changed. The structural forces favor whoever owns the data; the cyclical ones set the timing.
Strength bars are a research synthesis of dated public evidence: relative and directional, not spend, survey, or attribution data.
Playbook
Six plays follow from where the market is heading
Each play is something a marketer inside a filtration business can act on regardless of tooling, drawn from the moves already visible in the evidence.
Treat the cross-reference as a measured channel
Cross-reference searches are the highest-intent event in the category. Tying each lookup to the downstream reorder turns a catalog utility into the one channel whose value can be priced against revenue.
Defend part-number resolution against answer engines
If a language model answers the cross for a part number, the click never reaches the tool. Structured part data, schema markup, and licensed feeds keep the brand inside the resolution.
Budget for demand a parent used to provide
A spun-off or challenger brand inherits an installed base but not a funnel. Brand search, content, and attribution become new line items exactly when scrutiny on marketing spend is highest.
Close the co-op and distributor reporting gap
Spend flows through distributors and dealers, so brands rarely see sell-through by campaign. Blending co-op, distributor POS, and digital data into one view is the hardest measurement problem in the category.
Turn sensor signals into reorder prompts
Filter sensors and machine telematics can convert a filter-spent event into an automated, measurable reorder trigger. Whoever productizes that prompt owns the timing of the sale.
Reframe the price threat as a risk decision
Tariffs raise will-fit landed cost, but copies still win on sticker price. Filtration performance, warranty, and uptime evidence convert a price comparison into a total-cost argument the incumbent wins.
Method
How the research was built
Public filings, press releases, and live product pages captured on July 29, 2026. Every figure carries its source and capture date.
Four independent manufacturers or groups, three OEM genuine-parts programs, WIX as a group brand, Parker as its diversified parent, and the will-fit import category as a bookend.
SEC filings, investor materials, company newsrooms, regulator pages, and open product tools. No private, client, or platform data appears in this edition.
Map positions, the query-intent ladder, and signal strengths are analytical constructs (est.), not survey or spend data; reported third-party figures keep the reported label.
- Atmus Filtration investor relations FY2024 revenue of $1.67B and the roughly 80% aftermarket sales mix (reported). ↗
- Cummins newsroom The Atmus IPO of May 2023 and the completed split-off in 2024. ↗
- Donaldson Company FY2024 net sales of $3.5B and the Mobile Solutions segment structure (Form 10-K). ↗
- Shop.Donaldson The open cross-reference and e-commerce path of the most digitally mature independent. ↗
- Fleetguard The Fleetguard cross-reference tool and FIT filter-monitoring platform. ↗
- MANN+HUMMEL Group sales of roughly EUR 5B (reported) and the 2016 WIX Filters and Purolator acquisition. ↗
- WIX Filters The WIX online cross-reference, one of the strongest lookup experiences in the category. ↗
- Baldwin Filters The Baldwin cross-reference catalog and distributor-led aftermarket positioning. ↗
- Parker Hannifin newsroom The 2017 Clarcor acquisition of about $4.3B that brought Baldwin Filters into Parker. ↗
- Parts.cat.com The dealer-gated genuine-parts portal pattern the OEM programs share. ↗
- S&P Global Mobility The 2025 average age of US vehicles in operation, about 12.6 years (reported). ↗
- US EPA The heavy-duty NOx rule finalized in December 2022, effective model year 2027. ↗
- USTR Section 301 tariffs of 25% on many imported-parts categories; later escalation varies (est.). ↗
Independent analysis by Improvado based on public information captured in July 2026. Company names, product names, and marks belong to their respective owners. Directional scores and category conclusions are research synthesis, not audited performance data; figures marked est. are analytical estimates and figures marked reported are third-party reports.