Pulscar analysis of NHTSA public data, September 2026 Ford Focus: what 12,087 owner complaints, 29 recall campaigns and eight Ford programs say about the PowerShift years and what came after them Licensed CC BY 4.0. Queries run 19 September 2026. -------------------------------------------------------------------------- SOURCE FILES -------------------------------------------------------------------------- NHTSA ODI complaint flat files COMPLAINTS_RECEIVED_2015-2019, _2020-2024, _2025-2026 https://static.nhtsa.gov/odi/ffdd/cmpl/ (field dictionary: CMPL.txt) 777,040 distinct vehicle complaints (ODINO, product type V) received 1 January 2015 to 8 September 2026. Of those, 12,087 name MAKE = FORD and MODEL = FOCUS, and 6,372 of those carry a model year of 2014 or later. SCOPE NOTE: the flat file treats each variant as its own model. FOCUS ELECTRIC (108 complaints), FOCUS RS (6) and FOCUS SVT (1) are NOT in the 12,087. Where the article discusses the Focus Electric or the RS it says so and uses recall documents, not these counts. NHTSA recall flat file FLAT_RCL.zip. NHTSA manufacturer communication flat files TSBS_RECEIVED_2015-2019, _2020-2024, _2025-2026 https://static.nhtsa.gov/odi/ffdd/tsbs/ (field dictionary: TSBS.txt) 322 distinct manufacturer communications name the Focus. NHTSA investigation flat file FLAT_INV.zip. Five Ford dealer letters and one owner letter were downloaded as PDFs from static.nhtsa.gov/odi/tsbs/ and read in full; they are listed below and in ford-programs.csv. -------------------------------------------------------------------------- CASE SENSITIVITY -- READ THIS BEFORE REPRODUCING ANY COUNT -------------------------------------------------------------------------- NHTSA changed how it stores complaint narratives during 2021. Narratives received through 2020 are entirely upper case; those received from late 2021 onward are mixed case. EVERY query below is case-insensitive. This matters more on this vehicle than on most: 390 of the transmission-symptom complaints were filed from 2022 onward, and an upper-case-only regex would have dropped almost all of them -- along with the entire oil-pressure cluster, which is a 2024-2026 story. -------------------------------------------------------------------------- COMPONENT COUNTS -------------------------------------------------------------------------- NHTSA's COMPONENTS field holds one or more component names separated by " | ". A complaint counts once for each distinct component family it names, so family counts sum to more than 12,087. POWER TRAIN 5,480 (45.3% of the file) ENGINE 1,974 STEERING 1,520 FUEL/PROPULSION SYSTEM 1,346 UNKNOWN OR OTHER 1,334 ELECTRICAL SYSTEM 1,155 STRUCTURE 586 VEHICLE SPEED CONTROL 520 FUEL SYSTEM, GASOLINE 367 ENGINE AND ENGINE COOLING 292 LATCHES/LOCKS/LINKAGES 279 This is the manufacturer-neutral field NHTSA assigns; it is not the owner's wording. -------------------------------------------------------------------------- MILEAGE AND RATE FIELDS -------------------------------------------------------------------------- Mileage: NHTSA's MILES field, rows with 0 < miles < 400,000 only. All Focus complaints: 8,674 of 12,087 report a usable mileage. Median 66,200; first quartile 38,298; third quartile 96,000. Model year 2014 and later: 4,175 usable mileages, median 56,081. Transmission-symptom subset: 1,022 usable mileages, median 45,000; 42.7% at or under 36,000; 62.4% at or under 60,000; 87.4% at or under 100,000. Oil-pressure subset: 118 usable mileages, median 74,500. All-complaint baseline (whole 777,040-row file): 392,935 usable mileages, median 70,000. Towed / crash / fire: NHTSA's TOWED, CRASH and FIRE flags, share = Y. All Focus: towed 6.46%, crash 2.78%, fire 0.48%, 281 injuries, 9 deaths. Transmission-symptom subset: towed 2.48%, crash 1.28%, FIRE 0.00% (zero of 1,410), 14 injuries, 0 deaths. Oil-pressure subset: towed 39.91%, crash 0.46%, fire 0.92%. All-complaint baseline: towed 5.11%, crash 4.37%, fire 1.82%. Read those two subsets side by side. The transmission that defines this car's reputation sits BELOW baseline on every safety flag. The 1.0L oil-pressure cluster carries the highest tow share in the file by a factor of six. That contrast is the article's central argument about which failure to treat as urgent. -------------------------------------------------------------------------- THE TRANSMISSION-SYMPTOM SUBSET (1,410) -------------------------------------------------------------------------- Membership, case-insensitive, full narrative: (transmission|clutch)[^.]{0,150}(shudder|judder|jerk|lurch|slip| hesitat|surge) OR the same pair in the other order. FALSE-POSITIVE READING. Thirty matches were drawn at random (seed 42) and read in full. All thirty describe a transmission or clutch symptom on the car in question. Measured false-positive rate 0 in 30. The subset is deliberately narrower than two looser queries that were also run and are reported in subset-sizes.csv: 5,247 complaints (43.4%) use a transmission, clutch, PowerShift or DPS6 word at all, and 410 (3.4%) name the PowerShift or DPS6 by name. The 1,410 is the set where an owner wrote both a component and a symptom, which is the smallest defensible definition. Phrase shares inside the subset are published in what-owners-say-about-the-dps6.csv. The headline ones: symptom returned or the repair did not hold ......... 493 (35.0%) clutch replaced or repaired ......................... 348 (24.7%) names a recall ...................................... 292 (20.7%) calls it dangerous or describes a near miss ......... 275 (19.5%) mentions a stall .................................... 197 (14.0%) transmission control module named ................... 178 (12.6%) transmission replaced or rebuilt .................... 164 (11.6%) software, reprogramming or calibration .............. 161 (11.4%) names a program, extension or bulletin .............. 150 (10.6%) coverage expired, denied or refused ................. 127 (9.0%) parts not available or back-ordered .................. 41 (2.9%) input shaft seal ..................................... 18 (1.3%) WHAT "35.0% DID NOT HOLD" IS AND IS NOT. It counts complaints containing a persistence word -- again, still, came back, recurred, second time, did not fix, repeatedly -- anywhere in the narrative. Some of those words refer to a repeated symptom rather than a repeated repair, and a complaint filed after a successful repair is rare by construction, because people who are satisfied do not file. It is a measure of what owners who complain say, not of how often a DPS6 repair fails. By model year: 2012 = 273, 2013 = 230, 2014 = 395, 2015 = 148, 2016 = 178, 2017 = 119, 2018 = 60. As a share of each model year's own complaints: 2012 8.7%, 2013 13.3%, 2014 16.1%, 2015 14.6%, 2016 15.1%, 2017 13.3%, 2018 7.2%. Owner-reported dollar figures: every $N in the subset with 100 <= N <= 15,000. 124 figures. Median $1,800, first quartile $833, third quartile $3,000, largest $12,000. THIS IS NOT A PRICE: owners who write a number are the ones who paid one, and one complaint can carry both a quote and an invoice. -------------------------------------------------------------------------- THE OIL-PRESSURE SUBSET (218) -------------------------------------------------------------------------- Membership: the narrative contains "oil pressure" or "oil pump", case-insensitive. By model year: 2015 = 13, 2016 = 35, 2017 = 62, 2018 = 104 -- that is 98.2% of the subset inside model years 2015-2018, which is the 1.0L population of recall 23V-905 and investigation EA26004. Towed 39.91%. Median odometer 74,500 miles across 118 usable readings. NHTSA's own PE25020 closing resume puts the average failure at "approximately 70,000 miles" and says "98% of the failures occurred prior to the 150,000 mile timing belt replacement interval." Our median is an independent number from a different field of the same public dataset, and it lands in the same place. The narrower query "timing belt" returns only 26 complaints. That is not evidence the belt is rarely the cause -- owners write down the warning light they saw, not the part a teardown found. The article says so. -------------------------------------------------------------------------- OTHER SUBSETS -------------------------------------------------------------------------- stall ............................................ 1,973 (16.3%) steering assist fault / loss of power steering ... 1,042 (8.6%), 592 of them model year 2012. No recall campaign in the file covers 2012 Focus power steering. Published as a cluster without a campaign, not as a defect finding. door latch ....................................... 1,004 (8.3%), 558 of them model year 2012 -- campaigns 16V-643 and 20V-331. evaporative canister or purge valve ................. 575 (4.8%) fuel tank ........................................... 492 (4.1%) coolant or overheating .............................. 419 (3.5%) engine seized, failed or replaced ................... 277 (2.3%) names a recall .................................... 3,389 (28.0%) names a class action, lawsuit or settlement ......... 229 (1.9%) -------------------------------------------------------------------------- RECALLS -------------------------------------------------------------------------- 29 distinct campaigns name the Ford Focus in the eleven-year file, with a combined population of 6,939,205 vehicles counted once per campaign. THAT TOTAL IS NOT A FOCUS NUMBER: NHTSA publishes one population per campaign across every make and model in it. No per-model split exists. NO CAMPAIGN IN THE FILE RECALLS THE DPS6 POWERSHIFT ITSELF. The three power-train campaigns that do exist are 13V-523 (2,456 vehicles, transmission control module), 18V-169 and 18V-845 (manual transmission clutch, 5,361 and 2,828) and 26V-376 (5,252, June 2026, manual transmission again). The PowerShift was handled by customer satisfaction programs and by litigation, not by a safety recall, and the article says so plainly. -------------------------------------------------------------------------- DOCUMENTS READ IN FULL (PDF) -------------------------------------------------------------------------- 14M02 Supplement #7, 4 May 2023 (MC-10237441-0001) -- DPS6 TCM extended warranty, 10 years / 150,000 miles, transferable; out-of-term one-time TCM replacement "through June 30, 2025"; owner refund offer "expires November 30, 2023"; "There is no expiration date for emergency repair refunds." 19N08, 14 August 2019 (MC-10164603-0001) -- DPS6 clutch warranty extension, 7 years / 100,000 miles, 2014-2016 Focus built 9 July 2013 to 5 November 2015; out-of-term window to 10 February 2020. 24N07, dealer letter 4 December 2024 (MC-11012235-0001) -- one-time canister purge valve replacement, 2012-2018 Focus built 1 August 2010 to 10 May 2018, US population 1,483,446, "15 years of service (unlimited miles) from the warranty start date", "Coverage is automatically transferred to subsequent owners." 25E05 owner letter, March 2026 (MC-11030920-0001) -- emissions recall, 2015-2018 Focus 1.0L: inspect the CPV, replace if not at the latest hardware level, reprogram the PCM; "Even if you had 24N07 performed on your vehicle, you still need to have 25E05 performed." 26L02, dealer letter 25 August 2026 (MC-11036546-0001) -- 1.0L Fox Classic timing belt and oil pump belt interval change, US population 146,945, new interval "100,000 miles or 6 years", with reimbursement criteria including DTCs P0016, P0017, P0018, P0019, P033F and documentation of the last three oil changes. Investigation resumes PE25020, EA26004, RQ23006 and PE23015 were read from the NHTSA investigation flat file. -------------------------------------------------------------------------- INVESTIGATIONS -------------------------------------------------------------------------- PE23-015 opened 20 Sep 2023, closed 14 Feb 2024, oil pump failure -- closed on recall 23V-905 RQ23-006 opened 15 Sep 2023, closed 20 Dec 2024, loss of motive power -- the recall query into 18V-735 and 19V-515 that produced 24N07 PE25-020 opened 18 Dec 2025, closed 31 Jul 2026, timing belt failure EA26-004 opened 31 Jul 2026, STILL OPEN, timing belt failure PE16-001 / RQ19-005 door latch failures, 2016-2022 EA02-022 / EA02-023 engine stalling and front suspension collapse, 2002-2004, on the previous-generation car -------------------------------------------------------------------------- LIMITATIONS -- READ BEFORE QUOTING ANY NUMBER -------------------------------------------------------------------------- 1. Complaints are self-reported, unverified and voluntary. They are not failure rates and cannot be turned into one, because NHTSA publishes no denominator. 2. Counts follow attention. This file is the clearest case in the series: the 2019 spike (2,177 complaints) sits on top of the door-latch recall 20V-331 and the DPS6 programs 19N07 and 19N08, all of which mailed letters. 3. The complaint file records no engine code and no transmission type. Every 1.0L-specific or DPS6-specific statement in the article rests on a document, not on these counts; where a subset lines up with a program's model years, the article says that is a line-up, not a proof. 4. Mileage, towed, crash and fire are the owner's own entries. The transmission subset's 2.48% towed flag sits against narratives that describe being towed far more often than that. 5. Model year in the complaint file is the owner's entry and is sometimes wrong; two complaints carry model year 9999. 6. Coverage is decided by VIN and build date, never by model year alone. Ford's own populations here are defined by assembly plant and build date, and both 14M02 and 24N07 cut across model years. -------------------------------------------------------------------------- FILES -------------------------------------------------------------------------- complaints-by-model-year.csv model_year, focus_complaints, power_train_component_n, power_train_pct, transmission_symptom_complaints, transmission_symptom_pct, mileages_reported, median_miles. what-owners-say-about-the-dps6.csv phrase_group, complaints, pct_of_1410. oil-pressure-complaints.csv model_year, oil_pressure_or_pump_complaints, towed_pct, mileages_reported, median_miles. ford-programs.csv program, date, subject, focus_applicability, terms, status_september_2026. recall-campaigns.csv campaign, report_date, component, potentially_affected_whole_campaign, focus_model_years, consequence. subset-sizes.csv subset, complaints, pct_of_12087. No personal data is published: no VIN, no city, no dealer name, no owner free text beyond short quoted snippets in the article itself.