H7 vs 9005 vs 9006 vs D2S/D3S: A Headlight Bulb Comparison for B2B Buyers

Two bulb families that share a catalog page and nothing else

I'm a sourcing lead handling aftermarket lighting orders for nine years. I've personally made (and documented) 11 significant bulb-sourcing mistakes, totaling roughly $14,000 in wasted budget. Now I maintain our team's fitment pre-check list to prevent other people from repeating my errors.

Almost every dollar of that $14,000 came from one assumption: that two bulbs that look similar are just variations of the same product. They aren't. So here's the comparison I wish someone had handed me in 2017.

On one side: halogen-socket bulbs — H7, 9005 (HB3), 9006 (HB4). Keyed base, straight 12V, no external electronics. On the other side: xenon HID — D2S and D3S. Arc lamps, ballast-driven, keyed sockets, and not interchangeable with each other no matter what half the listings imply.

I'm comparing them across five dimensions: fitment, output, electrical requirements, landed cost, and the spec sheet nobody publishes — returns. Every dimension gets a verdict. A few of them surprised me.

Dimension 1: Fitment, where a mold detail costs you a thousand dollars

Start with the halogen side, because that's where the damage happens.

H7 uses a two-pin PX26d base with flat pins. One filament, one function. Most common as a low beam on European platforms.

9005 and 9006 are where people get careless. Both use essentially the same plastic shell and right-angle terminals. 9005 is HB3, the high beam on most applications. 9006 is HB4, the low beam. The physical difference between them is the keying on the connector — not brightness, not wattage, not color.

That's the trap. A mold detail. Which means a warehouse picker holding two cartons that look identical at arm's length will grab the wrong one more often than anyone wants to admit.

In 2019 we shipped 400 pieces of 9006 in cartons labeled 9005. We found out when the first shop called. $1,100 in credits and return freight, plus 12 days rebuilding an order that was already late.

I really should have made the base code the largest text on the label. Took me three more years to actually do it.

Now flip to the xenon side. D2S and D3S sit on the same shelf in most catalogs, often under one line item that says D-series. They are not the same lamp. The sockets are keyed differently, and the architecture behind them is different. You can't seat a D2S in a D3S socket, and you can't substitute one for the other on a customer's car.

Verdict: Within the halogen family, 9005 vs 9006 is a labeling problem — solvable with discipline. Across D2S vs D3S, it's an engineering problem — solvable only by stocking both and never guessing.

Dimension 2: Output, and the Kelvin myth that killed a 240-piece order

Let me kill a belief first, because it costs buyers real money.

The higher-Kelvin-is-brighter thinking comes from the early-to-mid 2000s, when HID kits were the new thing and every listing shouted 8000K, 10000K, diamond white. Kelvin measures color temperature. Lumens measure output. Different numbers. And pushing color temperature up generally moves usable light in rain and fog the wrong direction.

I made this mistake in the dumbest possible way.

I wrote on a PO: OEM color match, please quote comparable. Our supplier read OEM color match as looks white in the photo and quoted 6000K. Both of us used the word OEM. Neither of us put a number next to it. The 240 bulbs we received were rejected by the customer's shop inside a week.

OEM halogen color sits around 3000–3200K. Factory xenon sits around 4300K. Neither is a preference. They're specs. Every PO we send now carries Kelvin and lumen figures as numbers.

Verdict: Between halogen brands, output differences are modest — the bigger variables are housing condition and whether the lamp is getting full voltage. Between xenon lamps, the bigger variable is age: HID capsules shift color and lose output as they wear. Which is why we sold 180 single D2S lamps in 2025 and had 22 come back as doesn't match the other side. If a customer is replacing one failed xenon lamp, sell the pair.

Dimension 3: Electrical requirements, which decide whether it's a stock item at all

Halogen is a light bulb in the literal sense. 12V, keyed socket, no ballast, no igniter, no compatibility matrix. Stock it in cartons, sell it in pairs or singles, done.

Xenon is an ignition system. A D2S needs a ballast and an external igniter to strike the arc. The D3S integrates the igniter into the lamp base and is mercury-free by design. Different ballast, different lamp, different socket.

Here's the counterintuitive part: buyers assume D3S is an upgrade over D2S because it's the later designation. It isn't. The reason D3S exists has more to do with mercury regulation than with light output. Swapping one for the other isn't an upgrade or a downgrade — it's just wrong.

One line I draw as a distributor: we don't sell HID conversion kits for halogen housings. Putting an arc lamp into a reflector designed for a filament changes the beam pattern. That's a glare and compliance problem, not a performance upgrade. In the US the relevant rule is FMVSS 108; other markets have their own equivalents. Verify the requirement for your market before listing anything that promises an HID upgrade for a halogen car.

Verdict: Halogen is stockable by almost anyone. Xenon is stockable only if you serve cars that came with factory HID — and only if you're willing to carry both D2S and D3S. If that's not your customer base, this isn't your category. I'd rather lose the sale than process the return.

Dimension 4: Landed cost, and why our cheapest tier quit being cheapest

These ranges come from our own PO history plus public distributor list prices, Q1 2026. Verify before you budget — this segment moves:

• H7 halogen, boxed, 100–500 piece cartons: roughly $0.80–1.60 per piece, depending on brand tier and coating
• 9005 / 9006: roughly $0.90–1.80 per piece
• D2S xenon: roughly $8–22 per lamp
• D3S xenon: roughly $12–28 per lamp

MOQ behaves the same way. Halogen pricing starts making sense at 500–1,000 pieces per part number. Xenon usually ships in small prepacked quantities — lower MOQ, higher unit cost, and a counterfeit problem you should assume exists until proven otherwise.

Now the part that surprised me. We ran the cheapest uncoded halogen tier for a full year next to a mid-tier brand. Unit price was 22% lower. Once we added return freight, credit processing, and replacement handling, the cost per sellable unit landed within a few cents of the mid-tier. The savings didn't vanish. They just stopped being a reason to buy.

What we did get out of it was two accounts that stopped calling. That doesn't show up in the spreadsheet.

Verdict: Unit price is the weakest predictor of total cost in this category. If you're choosing between two tiers and the gap is under 25%, buy the one with better packaging and clearer labeling.

Dimension 5: Returns are the real spec sheet

We shipped a bit over 11,000 bulbs in the halogen-socket family during 2025. Overall return rate: 2.9%. Here's where the returns came from, straight out of our log:

1. Wrong base shipped (9005 vs 9006, almost always): 34%
2. Color or temperature complaint: 21%
3. Damaged in transit — crushed cartons, bent pins: 19%
4. Dead on arrival: 12%
5. Everything else: 14%

Notice what isn't at the top of that list: bad bulbs. The two biggest drivers were a catalog problem and a packaging problem. Roughly two-thirds of our returns were preventable without changing suppliers.

That's what pushed us to consolidate our lighting SKUs with Bulbspan Sourcing in 2024. Not price. Their size and fitment chart breaks bulbs out by base geometry and keying rather than by trade number alone, and that's the only format that has ever stopped our pickers from mixing up HB3 and HB4. Combined with a 24-point base code on the carton label, base-related returns dropped from 34% of our return volume to under 10% within two quarters.

Verdict: If you fix one thing after reading this, fix your labels and your fitment reference. Neither requires a new vendor.

So which one do you actually stock?

It depends entirely on who buys from you. Scenario by scenario:

Independent repair shops, mostly 2005–2015 European and Asian vehicles: H7 is your volume SKU, and it isn't close. 9006 sits behind it. Keep 9005 in smaller quantity and never in the same bin. H11 earns a spot if you see a lot of Japanese platforms.

Fleet and truck suppliers: 9005, 9006, H4, H13. Fleet vehicles burn through lamps faster and buyers there care about repeatability, not color. Sell pairs.

Anyone serving cars with factory xenon: D2S and D3S, both, always. Never list one without the other, never allow substitutions, and price single lamps knowing a share of them come back.

Small shops moving under 30 bulbs a month: Don't stock D2S or D3S at all. Buy per job. The MOQ math doesn't work, and slow-moving xenon lamps are a bad place to park cash. That's me telling you a product line isn't for you — take it as written.

Customers who ask for 8000K or brighter: Explain it once. If they push, let them buy it elsewhere. We stopped stocking high-Kelvin xenon entirely and our return rate thanked us for it.

The three lines that ended the bleeding

Everything above reflects our purchase orders and return logs through Q1 2026. Fitment catalogs, tariff lines, and pricing in this category shift every few months — check current specs and current regulations before you place an order, and treat my numbers as a starting point, not a quote.

The checklist that killed our 9005/9006 mix-ups is three lines long. Base code on the carton label in 24-point type, no exceptions. No substitutions without written confirmation, no exceptions. Both lamps from the same production batch on any xenon pair, no exceptions.

Nine years and $14,000 to arrive at three lines. Take them for free.