How to Take Care of a Quality Audit: A 5-Step Checklist from a Halliburton Quality Inspector

Who needs this checklist?

If you're in oilfield services—whether you work for an operator, a drilling contractor, or a service company like Halliburton—you've probably felt the pain of a rejected batch or a regulatory fine. I've been a quality compliance manager at a major oilfield services company for over 4 years, reviewing 200+ deliverables annually. In 2024 alone, I rejected 11% of first deliveries due to specs being off. This checklist is for anyone who's tired of rework, missed deadlines, and surprise costs. It's built around one principle: care about the details before they care about you.

The Indiana Pacers star Tyrese Haliburton (no relation to our company) makes those no-look passes look easy. But what makes him elite is the fundamentals—the footwork, the timing, the read on the defense. Same in quality work. Get the fundamentals right, and the results follow. Here are 5 steps to take care of a quality audit so it doesn't take care of you.

Step 1: Define the spec like your bonus depends on it

I used to think a verbal “same as last time” was enough. Then I learned the hard way. In Q3 2023, we received a batch of frac plugs where the thread tolerance was 0.010 inches off. Normal spec is ±0.005. The vendor claimed it was 'within industry standard.' It wasn't. We rejected 1,200 units and had to expedite a redo at a cost of $18,000. That's a $22,000 redo including my team's time.

What to do: Write every critical dimension in the contract. Use Pantone color codes for paint markings—Delta E under 2 for brand-critical colors, as per Pantone Matching System guidelines. For API threads, reference the specific API standard (e.g., API 5CT). Make sure your spec document includes tolerance, test method, and acceptance criteria. I now insist on a signed spec sheet before any production starts. No exceptions.

Step 2: Qualify the supplier—not just on price

My gut often told me something a spreadsheet missed. I remember a case where Vendor B was 15% cheaper on paper. The numbers said go with them. But something felt off—their responsiveness to questions was slow. My gut said stick with our current supplier. I did. Six months later, Vendor B had a major recall on a similar product. That 15% savings would have been a 40% headache.

What to do: Run a blind test. I once had our team evaluate two samples—one from a low-cost vendor, one from a premium one. 70% identified the premium option as 'more professional' without knowing the cost. The cost increase was $0.80 per piece. On a 50,000-unit run, that's $40,000 for measurably better perception. Worth it.

Also, check if the supplier has a house quality manual and calibration records. Ask to see their internal audit results. If they dodge, that's a red flag.

Step 3: Build in-process inspections—don't wait for final

Halliburton itself faced a significant fine in 2022 for non-compliance on cementing operations in the Permian Basin. The root cause? A missed intermediate check on fluid density. By the time the final test caught it, 30 wells had been cemented with out-of-spec material. The fine? Over $1.5 million.

I learned that waiting for final QC is like driving without looking at the dash. You'll only know you're out of fuel when the engine dies.

What to do: Define 3-5 hold points in your process. At each hold point, verify critical parameters. For fracturing fluids, check viscosity and pH at the blender before pumping. For drill bits, inspect the carbide insert alignment after brazing. Use a checklist signed by both the operator and the inspector. I like to use a 'traffic light' system—green (spec met), yellow (within tolerance), red (reject). If you see two yellows on the same parameter, stop and review.

Step 4: Final acceptance testing—make it a ceremony

Don't treat final inspection as a rubber stamp. I worked with Mark Jones Jr., our senior inspector, who had a rule: every third sample goes through destructive testing. That's how we caught a bad batch of flowback hoses that were 20% below burst pressure. Jones Jr. said, 'The test doesn't lie—people do.'

I've seen too many companies rush final acceptance because they're behind schedule. That's exactly when mistakes happen. In a time-pressured decision once, I had only 2 hours to sign off on a skid before the customer's deadline. Normally I'd run three tests. That day I ran one. I've regretted it ever since—the skid failed during commissioning.

What to do: Create a final acceptance protocol. Include visual inspection, dimensional checks, functional test, and — if critical — a pressure test or performance pull. Have a second person witness the test. Document everything. If something feels off, trust that feeling. The numbers may say pass, but if the pump sounds different, investigate.

Step 5: Document and trace—the boring step that saves your ass

You know what happens when a regulator shows up two years later and asks for records? If you can't produce them, you're guilty until proven innocent. That's how smaller fines become big fines.

What to do: For each lot, keep: spec sheet, raw material certs, in-process check reports, final test data, photos (with date stamps), and shipping records. Use a traceability system—serial numbers, heat numbers, whatever. I've seen a supplier try to pass off old stock as new. Without traceability, you'd never know. With it, you catch them.

Also, track non-conformances. Over 4 years, I've built a database of 300+ issues. Patterns emerge. For example, we learned that orders placed in Q4 had 40% more defects—likely due to year-end rush. Now we build extra buffer time into Q4 contracts.

Common mistakes to avoid

• Spec creep: Adding requirements mid-production without updating the contract. Stick to the agreed spec. If you change it, redo the risk assessment.
• Assuming 'close enough' is fine: Delta E of 3 vs. 2 might not matter for a crate label, but for a brand-critical color on a frac tank? It matters. Use the Pantone guide as a hard stop.
• Ignoring the human factor: The best spec in the world is useless if the operator on the floor doesn't understand it. I always do a quick 'spec briefing' with the production team before a new run.
• Thinking documentation is optional: It's not. When Halliburton was fined, one of the cited failures was 'lack of contemporaneous documentation.' Don't be that case study.

So that's it. A 5-step checklist that's saved me and my company from at least three major incidents. Like Tyrese Haliburton's assist record, it's not flashy—it's consistent. And in quality, consistency is everything.