Austin Product Liability Lawyer

Do not throw it away, do not let anyone take it

This is the only part of this page that is urgent, so it goes first.

In a car wreck the evidence is the crash report and the medical records, and both of those keep. In a product case the evidence is the product. If it goes in a dumpster, gets scrapped with the vehicle, or leaves with somebody else's investigator, there is usually no case left to bring. An engineer cannot inspect a photograph of a battery cell. A jury cannot look at the serial number you meant to write down.

So, before anything else:

• Keep the object. Burned, melted, snapped, whatever condition it is in. Put it in a garage, a storage unit, a plastic bin, anywhere dry and out of the way. Do not clean it, do not reassemble it, do not try the repair yourself.
• Keep the packaging, the manual, the charger, and the receipt. The charger matters enormously in a battery fire, because the first thing a manufacturer says is that you used the wrong one. The receipt or the order email is how we prove the date of sale, and the date of sale controls a deadline explained further down this page.
• Photograph the serial number, the model number, the lot or batch code, and any date stamp. Take them in daylight, close, and in focus. On a tire the code you want is the DOT number on the sidewall. On a vehicle it is the VIN on the dash and the door jamb. On an appliance it is usually a plate on the back or underneath.
• Do not hand it to the fire investigator, the insurance adjuster, or the store, and do not sign anything that lets them take it. They will ask, and the request sounds reasonable and procedural. Once the item is in the other side's custody you have lost control of the inspection, and sometimes the item. If a public fire department needs the item for an origin-and-cause investigation, that is a different situation and one we can work with. Call before you agree.
• Write down what happened while it is fresh. What you were doing, how long the product had been in use, whether it had done anything strange before, who saw it. Memory for this kind of detail degrades in weeks.

If the fire or the failure happened at a rental property, there is a second clock: the landlord's insurer will have a cause-and-origin expert out fast, and debris gets hauled. A short preservation letter from a lawyer, sent in the first days, is often the only thing that stops a scene from being cleared.

Three kinds of defect, and which one your case is

Texas puts every claim about a defective product into one chapter of the code and then divides it three ways. Chapter 82 of the Civil Practice and Remedies Code governs a products liability action, and § 82.001(2) defines that phrase broadly on purpose: any action against a manufacturer or seller for damages allegedly caused by a defective product, whether it is pleaded as strict liability, negligence, misrepresentation, breach of express or implied warranty, or any other theory. That definition is why "we'll just plead it as negligence" does not get anyone around the chapter. The provisions follow the claim.

Inside that, the three theories are:

1. Design defect. The product was built exactly as intended and the intention was dangerous. Every unit that came off that line has the same problem.
2. Manufacturing defect. The design was fine. Your unit came out wrong.
3. Marketing defect, better known as failure to warn. The product is what it is, and the risk was one the maker knew about and did not tell you about in a way that would have made a difference.

Which one applies changes the proof, the cost, and often the answer to whether there is a case. Sorting it early is the single highest-value hour in one of these files.

Design defect and the alternative that has to exist

This is where most Texas product cases end, and people are usually shocked by it.

Under § 82.005(a), if you allege a design defect, you carry the burden of proving by a preponderance of the evidence two things: that there was a safer alternative design, and that the defect was a producing cause of the injury. The alternative design is not a nice-to-have. Without it the claim fails as a matter of law, no matter how badly the product performed.

Section 82.005(b) then defines what counts. A safer alternative design is a design other than the one actually used that in reasonable probability would have prevented or significantly reduced the risk of the injury without substantially impairing the product's utility, and that was economically and technologically feasible at the time the product left the manufacturer's or seller's control, using existing or reasonably achievable scientific knowledge.

Read that as four separate hurdles, because that is how it gets litigated:

• It has to be a design, not a wish. "They should have made it safer" is not an alternative design. "They should have used a cell with a ceramic-coated separator and a pressure-relief vent, here is the part, here is the supplier, here is what it cost in that model year" is.
• It has to still work as a product. A space heater that cannot produce heat is safer and useless. Defendants win on this more often than people expect.
• It has to have been feasible then, not now. The question is the state of the art on the date that unit left the maker's control. A safety feature that became standard four years later does not automatically qualify.
• It has to have made a difference to your injury. Producing cause, tied to the specific mechanism that hurt you.

Practically, that means a design case needs an engineer before it needs anything else, and the engineer has to be willing to build and defend an alternative. That is the cost driver discussed at the bottom of this page. It is also why we will sometimes tell you a claim that feels obvious is not a design case at all, and is better brought as one of the other two.

Section 82.005 does not apply to toxic or environmental tort causes of action, or to drugs and devices regulated under the federal Food, Drug, and Cosmetic Act. Those run on different rules.

Manufacturing defect: the one that came off the line wrong

A manufacturing defect claim says the product deviated, in its construction or quality, from the manufacturer's own specifications or from otherwise identical units, and that the deviation made it unreasonably dangerous and produced the injury. Texas courts have described it in essentially those terms for years; Ford Motor Co. v. Ledesma, 242 S.W.3d 32 (Tex. 2007), is the usual reference point.

Here is the part that matters for screening: a manufacturing defect claim does not require a safer alternative design. Section 82.005 by its own terms applies to design defect allegations. If your unit is different from the other units, you do not have to redesign anything. You have to show the deviation.

That is a very different evidentiary job. Instead of an engineer proposing a better product, you need the manufacturer's own specification, the quality-control records for that plant and that run, and an examination of your unit showing where it departed. It is often cheaper. It is also often more fragile, because the deviation lives in the object, and the object is the thing people throw away.

There is one more asymmetry worth knowing. The government-standards presumption described below, the one that gives a compliant manufacturer a head start, does not apply to manufacturing defect claims at all under § 82.008(d). A defendant that leans on its federal certification has less to lean on when the argument is that this particular unit was not built to the standard it certified.

Failure to warn, and what a warning label has to say

A marketing defect claim is about information. Texas courts, following American Tobacco Co. v. Grinnell, 951 S.W.2d 420 (Tex. 1997), generally look for: a risk of harm inherent in the product or arising from its intended or reasonably anticipated use; that the supplier knew or should have known of the risk when the product was marketed; that the warning or instruction given was defective or absent; that the absence made the product unreasonably dangerous; and that the failure was a producing cause of the injury.

Two things decide most of these.

Was the risk obvious? There is no duty to warn that a knife cuts. There is a real question about whether an ordinary buyer understands that a lithium-ion pack charged on a cheap third-party charger can go into thermal runaway with no warning and burn hot enough that water makes it worse. Obviousness is judged from the position of an ordinary user, not an engineer.

Would a warning have changed anything? Texas applies a presumption that a user would have read and heeded an adequate warning if one had been given. That presumption is rebuttable, and defendants attack it with evidence about what you actually did, whether you read the manual, and whether you had been warned some other way. It helps. It does not decide the case.

A warning also has to be adequate in a practical sense. Buried on page 34 of a manual nobody opens, printed in six-point grey, or written only in English on a product sold heavily to Spanish-speaking households, are all arguments about adequacy rather than about whether words existed somewhere.

Fifteen years from the date of sale, whatever happened to you

Two clocks run in a Texas product case and they are measured from different events.

The first is the ordinary personal-injury limitations period: two years from the day you were injured, under § 16.003(a). If the fire was on a Tuesday, that Tuesday two years later is the outside edge, and waiting until month twenty-three is how good claims turn into malpractice questions.

The second is the one that surprises people. Under § 16.012(b), a products liability action must be commenced before the end of 15 years after the date of the sale of the product by the defendant. This is a statute of repose, and the distinction from a limitations period is not academic. A limitations clock starts when you are hurt. A repose clock starts when the product was sold, and it does not care that you had not been hurt yet. A sixteen-year-old table saw can take your hand off on a Saturday and the claim against the manufacturer can already be barred on the Sunday.

Two exceptions are written into the statute, and they are narrow:

• § 16.012(c). If the manufacturer or seller expressly warranted in writing that the product has a useful safe life longer than 15 years, the period runs for the number of years warranted instead. This is one of several reasons to keep the paperwork.
• § 16.012(d). The repose period does not apply to a personal-injury or wrongful-death claim where the claimant was exposed to the product within 15 years of first sale, the exposure caused the disease at issue, and the symptoms did not manifest within that window to a degree and duration that would put a reasonable person on notice of an injury. This is the latent-disease carve-out.

Section 16.012 also applies only to the sale of a product, not to a lease, and it does not extend any other limitations period.

So the first date we ask you for is not the date of the accident. It is the date of purchase. If you bought it new and you still have the order confirmation, bring that to the consultation. If it came with a house or a used vehicle, we may have to trace the first sale, and that trace sometimes ends the inquiry.

The store sold it, the manufacturer made it, and only one of them is usually liable

People often want to sue the store, because the store is nearby, has a Texas address, and is the only entity in the transaction they ever spoke to. In Texas that usually does not work, and § 82.003 is the reason.

A seller that did not manufacture the product is not liable for harm caused by that product unless the claimant proves one of seven specific things: that the seller participated in the design; that the seller altered or modified it and the harm came from that; that the seller installed it on another product and the harm came from the installation; that the seller exercised substantial control over the warnings or instructions and those were inadequate; that the seller made an incorrect factual representation about the product that the claimant relied on; that the seller actually knew of the defect when it supplied the product; or that the manufacturer is insolvent or is not subject to the jurisdiction of the court.

That last one, § 82.003(a)(7), is the exception that does real work. A large share of the products that injure people in Austin were made by an entity with no U.S. presence, no registered agent, no assets you can reach, and a seller name that changed three times on the marketplace listing. When the manufacturer cannot be brought into a Texas court, the seller's protection falls away. Establishing that is its own project, and it is one of the first things we look at on an imported-goods case.

Running the other direction, § 82.002(a) requires a manufacturer to indemnify and hold a seller harmless against loss arising out of a products liability action, except for loss caused by the seller's own negligence or misconduct. That statutory duty is a large part of why a case that starts with a retailer named as a defendant tends to end with the manufacturer standing in the middle of it, funding the defense and writing the check.

When the product complied with a government standard

Expect this defense in writing within weeks. Under § 82.008(a), a manufacturer or seller gets a rebuttable presumption of no liability if it establishes that the product's formula, labeling, or design complied with mandatory federal safety standards or regulations in effect when the product was made.

Rebuttable is the operative word. Section 82.008(b) says a claimant defeats the presumption by establishing that the applicable federal standards were inadequate to protect the public from unreasonable risks of injury or damage, or that the manufacturer withheld or misrepresented information material and relevant to the federal government's determination. Subsection (c) sets up a parallel presumption for products that went through pre-market government approval, with the same two ways out. Subsection (d) exempts manufacturing defect claims from both presumptions entirely.

Section 82.007 does the same job for pharmaceutical warnings that matched FDA-approved labeling, with its own list of five ways to rebut it, including that the defendant withheld or misrepresented material information from the FDA.

The short version for a client: "we passed the federal test" is a real advantage for the other side and it is not the end of the conversation. Federal minimum standards are floors, they are frequently older than the technology they govern, and what a manufacturer told the agency is discoverable.

Checking the recall databases yourself

You can do this today, for free, without a lawyer, and you should. Three federal databases cover most consumer goods.

CPSC — everything that is not a car or a drug. cpsc.gov/Recalls is searchable by product name and brand. CPSC published 420 recall announcements with a 2025 recall date (CPSC Recalls REST Web Service, queried 2026-09-05). Of those 420 records, 144 mention fire or burn and 45 mention lithium somewhere in the title, hazard, or description — that keyword count is ours, from the CPSC feed, and CPSC does not publish it that way.

NHTSA — vehicles, tires, car seats, and vehicle equipment. nhtsa.gov/recalls takes a VIN and tells you whether there is an open, unrepaired recall on that specific vehicle. Use the VIN, not the year and model, because recalls are usually limited to production ranges. Recall repairs at a franchised dealer are free.

FDA — medical devices and drugs. The device recall database is at accessdata.fda.gov, and the adverse-event reporting system, MAUDE, is at accessdata.fda.gov/scripts/cdrh/cfdocs/cfMAUDE. Read MAUDE with FDA's own warning in front of you. FDA states plainly that "the submission of an MDR itself does not necessarily demonstrate that the device caused or contributed to the adverse outcome or event," and that MDR data "is not intended to be used either to evaluate rates of adverse events, evaluate a change in event rates over time, or to compare adverse event occurrence rates across devices." The volume is enormous — 2,888,003 device reports carried a 2025 date received (openFDA, data last updated 2026-08-18) — and volume is not proof of anything. What MAUDE is genuinely good for is finding out whether anyone else described the same failure mode you experienced, in the same words, before it happened to you.

A recall does not decide your case

A recall helps. It is not a verdict, and clients who arrive holding a recall notice are often surprised by how much work is still ahead.

What a recall gives you: notice that the manufacturer knew, a documented description of the failure mode written by the people who built the thing, a population of affected units, and frequently an agency file to request. On a failure-to-warn theory, the date the manufacturer first learned of the problem is often the whole fight, and the recall paperwork is where that date lives.

What a recall does not give you: proof that your unit had the defect, proof that the defect caused your injury, or a safer alternative design. Recalls are also frequently voluntary and, in the motor vehicle world, overwhelmingly manufacturer-initiated. Of the 997 NHTSA recall campaigns with a 2025 report-received date, 878 were manufacturer-initiated; 107 were influenced by NHTSA's Office of Defects Investigation and 12 by the Office of Vehicle Safety Compliance (computed from NHTSA's ODI Recalls Flat File, pulled 2026-09-05).

And the absence of a recall proves nothing at all. Most defective products are never recalled.

Vehicle components: tires, airbags, seat belts, fuel systems

A large share of Texas product cases start as ordinary crash cases. The wreck happens, somebody is at fault for the collision, and then a second question appears: the crash was survivable and the occupant did not survive it, or the injuries do not match the forces involved. That is a crashworthiness question, and it is aimed at the vehicle rather than at the other driver.

The scale of the component world, from NHTSA's own recall record for 2025 (ODI Recalls Flat File, pulled 2026-09-05, counting distinct campaigns by report-received year):

A campaign can carry more than one component code, so the component rows overlap and do not sum to the totals above them.

Two Texas figures size the problem, both from NHTSA's Fatality Analysis Reporting System for 2024, filtered to Texas. Of 3,774 fatal crashes in Texas that killed 4,160 people, 822 — 21.8 percent — involved at least one vehicle coded as having rolled over. Rollover is where roof strength, door latches, and seat belt retention get tested, and it is the crash mode where a survivable event most often is not survived. And among 2,502 Texas passenger-vehicle occupant fatalities, 1,187 — 47.4 percent — were coded as wearing a shoulder and lap belt, with 1,010 coded as no restraint used or not applicable and 274 unknown.

Be careful with that last number, and we are careful with it in a case. A belted occupant who died did not necessarily die because the belt failed. Most of the time the forces were simply beyond what any restraint could manage. The figure tells you how often the question is even worth asking, which is more often than people assume.

Where these cases actually come from, in our experience: tread separation on a tire whose DOT date code shows it was years old when it was sold; an airbag that did not deploy in a frontal impact severe enough to trigger it, or deployed with the wrong force; a seat belt that unlatched or spooled out under load; roof crush in a rollover; and post-collision fuel-fed fire. Every one of those depends on the vehicle being preserved. If your car is at a storage yard accruing fees, tell us before it goes to salvage. Once it is crushed the case is over.

This page does not deal with who caused the collision, comparative fault at an intersection, or how a crash report is read. If that is your question, start with our Austin car accident page, and if you were on a motorcycle, the Austin motorcycle page. A component claim can run alongside either one.

Batteries, e-bikes, scooters and things that catch fire

Lithium-ion is where this practice area has moved, and Austin has the exposure: dockless scooters downtown, delivery e-bikes on Rainey and along Riverside, and a lot of apartments where the only place to charge is an entry hallway.

CPSC published its most recent analysis in April 2026, covering 2017 through 2024. The estimates below are national and come from NEISS, CPSC's hospital sample:

CPSC is aware of 533 reported fatalities across these products from 2017 through 2024, rising from 5 in 2017 to 135 in 2024. Within that, e-bikes account for 310 fatalities, of which 19 were associated with 13 lithium-ion battery fire incidents; e-scooters account for 206, of which 15 were associated with 11 battery fire incidents; hoverboards account for 17, of which 11 were associated with 5 battery fire incidents. CPSC's own note: reporting for 2023 and 2024 is ongoing and counts may change.

The number that tells you the most about these cases is not in the injury tables. CPSC field staff completed 531 in-depth investigations of micromobility incidents across those eight years, and fire was the most common problem reported in all three product families — 88 of 189 e-scooter investigations, 173 of 197 hoverboard investigations, and 106 of 145 e-bike investigations.

Read that as: when one of these devices seriously hurts somebody, it is usually because it burned.

A battery fire case has its own preservation list on top of the general one. Save the pack even if it is a fused lump, save the charger you actually used and any others in the house, save the wall outlet's circumstances if the fire department documented them, and get the fire department's incident number and cause-and-origin report. If the fire spread to a neighbor's unit or to the building, expect the property insurer's investigator to move within days, and expect them to want the pack.

The honest reason a manufacturer sometimes pays more than the medical bills

The damages discussion on most of this site lives on our compensation pages, and the ordinary categories work here the same way they work anywhere. This page goes deep on the one that shows up in product cases and almost nowhere else in ordinary negligence work: exemplary damages.

Chapter 41 of the Civil Practice and Remedies Code allows exemplary damages only when the claimant proves by clear and convincing evidence that the harm resulted from fraud, malice, or gross negligence. Clear and convincing is a higher standard than the preponderance standard that governs the rest of the case, and it is meant to be hard.

Product cases reach it more often than car wrecks do for a structural reason: manufacturers keep records. A distracted driver has a phone log and nothing else. A manufacturer has failure-mode analyses, warranty claim data, field reports, supplier correspondence, internal test results, and sometimes a cost estimate for the fix that was not made. When the documents show that a company knew a specific failure was occurring, quantified it, and decided the recall cost more than the claims, that is the fact pattern Chapter 41 was written for.

The cap is real and you should know it going in. Under § 41.008, exemplary damages are limited to the greater of two times economic damages plus non-economic damages up to $750,000, or $200,000. Anyone who describes an uncapped punitive recovery in a Texas product case is describing something that does not exist. More on the categories at our punitive damages page and, for a permanent injury, disability and disfigurement.

One more Texas rule that bites specifically here: proportionate responsibility under Chapter 33. The defense in a product case is almost always misuse — you overloaded it, you modified it, you ignored the manual, you charged it with the wrong charger. Your share of responsibility reduces your recovery, and if a jury puts you over 50 percent you recover nothing. Which is another reason we want the charger, the manual, and your account of what you were doing, early and accurate.

What these cases cost to prove, and why we say no to some of them

We work on contingency. You pay no fee unless there is a recovery, and the fee is a percentage of the recovery, agreed in writing before we start. That part is the same on every case this firm handles.

What is not the same is the case expenses, and product liability is the most expensive category of work in a plaintiff's office. The firm advances those costs. On a rear-end collision, advancing expenses means ordering records and paying a filing fee. Here it means:

• An engineer, retained early, to inspect the product and form an opinion on defect and, in a design case, on a specific alternative design. Not a general expert. Someone who works in the actual field — battery chemistry, tire construction, restraint systems, structural design.
• Destructive testing, done under protocol. Once the product is cut open it cannot be uncut. Both sides usually have to agree on a protocol, and everyone attends. That is engineer time, lab time, and lawyer time, all billed.
• An exemplar — an identical undamaged unit, bought and tested alongside yours to show what should have happened.
• A fire origin-and-cause expert on a burn case, separate from the engineer, because origin and defect are two different opinions.
• A medical or biomechanical expert where the fight is whether the defect, rather than the event, produced this injury.
• Deposition and document costs on a defendant that fights everything, sometimes overseas, sometimes through a corporate structure that has to be untangled before anyone can be served.

Then the expert has to survive a challenge. Texas admissibility runs through Rule 702 and E.I. du Pont de Nemours & Co. v. Robinson, 923 S.W.2d 549 (Tex. 1995), and in a product case the fight is about engineering method: was the testing protocol reproducible, was the alternative design actually built and tested or only drawn, does the failure analysis follow accepted practice in that discipline, and did the expert account for the other candidate causes. An engineer who cannot answer those questions gets excluded, and an excluded expert on a design case ends the case.

That is the honest economics, and it drives a hard rule we apply out loud. If the product is gone, if the sale was more than fifteen years ago, if the manufacturer is unreachable and the seller does not fit one of the seven exceptions, or if the expected recovery cannot carry six figures of expert cost, we will tell you that in the first conversation instead of signing you up and quietly letting the file age. Sometimes the better answer is a homeowner's or renter's claim, a warranty claim, or a CPSC report at saferproducts.gov that helps the next person. We would rather say that on day one.

Where the answer is yes, none of the above costs anything out of your pocket while the case runs.

A few practical things people ask before they call

Your immigration status has no bearing on a Texas injury claim and we do not ask about it. You do not need to come to our office; we handle intake and most of the case by phone, text, and email, and we come to you when a signature or an inspection needs a person. The consultation is in English or Spanish, whichever you prefer. If the product is bulky or burned, send photographs first and keep the item where it is until somebody tells you it is safe to move it.

Who would be handling this

Israel Medina, Founding Partner. Texas Bar No. 24130487. Juris Doctor, Baylor University School of Law. Admitted in Texas and Nevada. Member, Texas Trial Lawyers Association; Austin Bar Association; Travis County Bar Association; American Association for Justice. Israel handles cases from intake through trial and prepares each one as though it is going to a jury. Full bio.

Rikki Medina, Founding Partner. Texas Bar No. 24144569. Juris Doctor, Baylor University School of Law. Admitted in Texas and Nevada. Rikki runs case management and client communication, which on a product file means the preservation logistics, the inspection scheduling, and keeping you informed while the engineering work happens out of sight. Full bio.

Both attorneys practice in English and Spanish.

Sources

Every figure on this page comes from a federal database or a federal agency report. We pulled each one on 2026-09-05 and each is reproducible from the dataset itself. Where a number is national we have said so; the only Texas-specific figures on this page are the FARS 2024 rows, and they are labeled as Texas. Nothing on this page is an Austin-only figure, because no federal product-injury dataset reports at city level.

• CPSC, National Electronic Injury Surveillance System (NEISS) — hospital-sample injury estimates. NEISS draws from approximately 100 hospitals selected as a probability sample of all 5,000+ U.S. hospitals with emergency departments. CPSC's own framing: the system produces "national estimates of the number of injuries associated with (but not necessarily caused by) specific consumer products." An estimate associated with a product is not a finding that the product was defective. https://www.cpsc.gov/Research--Statistics/NEISS-Injury-Data
• CPSC, "Micromobility Products-Related Deaths, Injuries, and Hazard Patterns: 2017–2024," April 2026. All micromobility figures above. CPSC notes that fatality reporting for 2023–2024 is ongoing and counts may change. https://www.cpsc.gov/s3fs-public/Micromobility_Products-Related_Deaths_Injuries_and_Hazard_Patterns_2017-2024.pdf
• CPSC Recalls database and REST web service — 2025 recall count. https://www.cpsc.gov/Recalls
• NHTSA Office of Defects Investigation, Recalls Flat File — 2025 campaign counts, unit counts, and influenced-by breakdown, computed by distinct campaign number and report-received year. Cross-checked against NHTSA's 2025 Annual Report on Safety Recalls, published March 2026. https://www.nhtsa.gov/recalls
• NHTSA Fatality Analysis Reporting System (FARS), 2024 National file, filtered to Texas. FARS is a census of fatal crashes. It does not describe non-fatal crashes and it does not establish that any component failed. https://cdan.nhtsa.dot.gov/
• FDA MAUDE and the FDA medical device recall database — report volumes. FDA: "The submission of an MDR itself does not necessarily demonstrate that the device caused or contributed to the adverse outcome or event." https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfMAUDE/search.CFM · https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfRES/res.cfm

What is deliberately absent from this page. There is no count of defective-product lawsuits filed in Travis County, because no public dataset breaks civil filings out by product claim. There is no Austin-level injury figure, because NEISS is a national sample and does not report at city level. There is no dollar figure for any settlement or verdict this firm has obtained. There is no estimate of how many injuries are caused by defects, as opposed to associated with products, because no agency publishes that and constructing it would be guessing.

Common questions about a defective product

What people ask before they call us about a product that failed.

Related pages on this site

• A defective machine or tool on a job site
• A tyre, airbag or seat belt failure in a crash
• An e-bike or scooter battery fire
• A death caused by a defective product
• Burns, amputation and permanent injury

Talk to us about the product

This page is general information about Texas law. It is not legal advice, and reading it does not create an attorney-client relationship. Every case turns on its own facts. Past results do not guarantee future results.