Walk through any storage unit district and you will find unfinished projects. Shells on stands, parts in boxes, a folder of receipts, and a car that has been ninety percent done for eleven years. Ask what happened and the answer is almost never "I lost interest." It is almost always some version of "it cost twice what I thought and I ran out in the paint booth."
That failure is predictable and it is preventable, and it comes down to one structural misunderstanding: people budget restorations by adding up parts. Parts are the smaller half. The larger half is labour that cannot be rushed, applied to the phase that nobody prices properly in advance — metal work, bodywork and paint.
This guide is about where the money actually goes, in what order, and how to plan a project you can finish. It does not contain dollar figures, because they vary by region, by vehicle, by shop rate and by the standard you are chasing, and a number would be misleading. What it contains is the shape of the budget, which is consistent across almost every restoration, and the method for turning that shape into real numbers for your car.
The Core Arithmetic
Three facts govern everything.
Labour dominates. On a typical full restoration, labour is the majority of the total, often substantially. Parts feel like the expense because you can see them and price them online. The hours do not have a product page.
Bodywork is the largest single line item, by a wide margin. Metal repair, panel fitting, gap setting, filler work and block sanding are hundreds of hours of hand work on a thorough restoration. None of it can be accelerated by spending more, and the quality of the final paint depends almost entirely on how many of those hours were spent before any colour went on.
The small things add up to a large thing. Fasteners, clips, seals, gaskets, hoses, wiring, fluids, media blasting, plating, powder coating, consumables, abrasives, primer, shop supplies. Individually trivial, collectively a meaningful fraction of the total, and left out of almost every back-of-envelope estimate.
That third one deserves emphasis because it is the most consistently missed. A restoration consumes an extraordinary quantity of small items. A single door might need thirty fasteners, two seals, a weatherstrip, a check strap, a lock mechanism rebuild, a window channel, two glass runs and a handful of clips — none of which appear in anyone's mental budget.
The Phases, in Order of Cost
Metal work and bodywork — the largest
This is the phase that determines whether the project finishes. It includes cutting out and replacing corroded metal, repairing or replacing panels, fitting and aligning everything, setting panel gaps in bare metal, metal finishing, skim filler, and block sanding to a flat surface.
Three things make it expensive and unpredictable:
It is almost pure labour. Replacement steel is cheap. The hours fitting it are not.
The scope is unknown until you start. Corrosion extends well beyond what any inspection reveals, because it starts on the hidden side and progresses outward. Every restoration finds more rust than expected. This is not pessimism; it is the normal outcome, and the only defence is a contingency.
The finish standard changes the hours dramatically. A driver-quality car needs panels that look excellent from a few feet. A show car needs panels that are genuinely flat under direct light from any angle, and the difference is a large multiple of hours in block sanding.
Paint — second largest, and mostly preparation
Paint cost is dominated by preparation, not spraying. Primer, blocking, guide coats, sealer, more blocking, masking, and then colour and clear — followed by cutting and buffing, which on a high-standard job is itself substantial.
The material cost of a quality paint system is real but it is a fraction of the labour. A shop quoting a low price for paint is quoting less preparation, and preparation is the entire thing. Paint does not hide a wavy panel; it reveals it, in sunlight, permanently.
Interior — frequently underestimated
Seats, foam, upholstery, headliner, carpet, door panels, dash pad, gauges, trim, weatherstripping and insulation.
For a popular model, pre-sewn cover kits over new foam give an excellent result at a fraction of custom trim cost. For an unpopular one, or for anything bespoke, upholstery is skilled hand labour and priced accordingly. Headliners in particular are unforgiving and take a specialist a surprising amount of time.
Weatherstripping is the item people forget entirely, and it is both expensive and essential — a beautifully restored car that leaks will rot from the inside.
Drivetrain — predictable, and mostly parts
An engine rebuild, transmission rebuild or replacement, differential, clutch or converter, cooling package, fuel system and exhaust.
This is the phase people over-budget relative to its real share, because it is the phase with product pages and visible prices. It is also the most predictable: machine shop work has standard rates, parts have known costs, and the scope is discoverable by measuring.
Where it escalates is originality. A correct rebuild of an original carburettor, distributor and generator by a specialist costs considerably more than a complete modern injection system in a box — one of the several reasons a concours restoration frequently exceeds a restomod.
Brakes, suspension and steering — moderate and knowable
Complete replacement is normal on any long-stored vehicle, because lines corrode internally, hoses perish, and cylinders seize. Bushings, ball joints, tie rods, springs, dampers, and a full brake system.
Mostly parts, mostly available, mostly predictable. Budget for everything rather than for inspection, because on a car that has sat, the answer is almost always everything.
Wiring and electrical — moderate
A reproduction harness for a correct restoration, or a modern fuse panel and harness for a restomod. Plus gauges, senders, switches, lighting and however many relays the loads require.
Plating, polishing and brightwork — variable and surprising
Rechroming is expensive, slow, and done by a shrinking number of shops. Pot metal parts frequently need repair before plating. Stainless trim can be polished rather than plated, which is cheaper but still skilled hand work. Fasteners need the correct finish for a correct restoration, which means individual plating.
For a car with a lot of brightwork, this phase alone can rival the drivetrain.
Glass, seals and small parts — the long tail
Glass, rubber, fasteners, clips, badges, lenses, and the hundred items that hold everything together.
Where Projects Actually Go Wrong
The budget was set against the wrong car
The purchase price looked attractive, so the cheap shell got bought. But metal work is the most expensive labour in the project, so the cheap shell front-loads the expensive phase. Paying several times more for a structurally sound example almost always costs less in total. This is the single most consequential budgeting decision and it is made before the project starts.
Several phases are open at once
The classic failure pattern: engine out and partly rebuilt, floor pans half in, interior in boxes, trim in a corner, and nothing finished. Parts get lost. Parts get damaged. Bare metal flash-rusts and has to be cleaned again. Fasteners end up in a single bucket and nobody knows what came from where.
Finish each phase completely before starting the next. It is slower in feel and much faster in reality.
The scope moved
Six months in, somebody suggests a bigger engine, and now the tunnel needs modifying, the crossmember needs relocating, the driveshaft is wrong, the cooling package is undersized, and the brake proportioning needs rethinking. Each change looks small and each one cascades.
A written specification before teardown is what makes "no, that is not this project" a sentence you can say to yourself.
No contingency
The metal work will find more corrosion than the inspection did. A machine shop will find a cracked head. A reproduction panel will not fit as supplied and will need reworking. These are normal events, not disasters, and a project with no contingency treats each one as a crisis.
Everything was bought at once
Buying all the parts early feels like progress and is usually a mistake. You tie up capital, you accumulate parts that get damaged in storage, and you frequently buy things that the project no longer needs because a decision changed. Buy per phase, as you reach it.
Labour was priced optimistically
If you are paying a shop, get the figure in writing against a defined scope before starting. If you are doing it yourself, be honest about your hourly availability. A project needing four hundred hands-on hours, at four hours a weekend, is two years of continuous weekends with no holidays, no distractions and no setbacks.
Doing It Yourself Versus Paying a Shop
The right answer is usually a mix, and the division should be based on which phases reward skill and equipment most.
Worth doing yourself if you have the time and the inclination: teardown and documentation, which is slow and needs no skill but needs care; cleaning, degreasing and parts preparation; disassembly and cataloguing; mechanical assembly; brake, suspension and fuel line work; wiring, which is tedious rather than difficult; interior fitting of kit components; and reassembly.
Worth paying for: metal fabrication and welding on structure, unless you are genuinely competent, because this is both the most expensive phase and the one where mistakes are structural; paint, because a controlled environment makes an enormous difference to the finishing work afterwards; machine shop work, which needs equipment you will not buy; upholstery and headliners; plating; and alignment.
The honest trap in doing it yourself is time, not skill. Most of this work is learnable. What people underestimate is that a restoration is a thousand small tasks, and doing them in evenings and weekends while life continues takes years. A project that takes a shop six months takes an individual three to five years, and the shell is occupying space and deteriorating for all of it.
Building a Real Number for Your Car
Here is the method. It takes a weekend and it is the most valuable weekend in the project.
One: define the finish standard in writing. Driver, show or concours. This single decision changes the total by a multiple, because it changes the hours in bodywork and paint more than anything else.
Two: get the metal work and paint quoted by a shop that does this work, based on your inspection photographs and findings. Not a guess, not a forum opinion — a figure from somebody who will do it. This is the number that decides viability and it is the number almost nobody obtains in advance.
Three: build a parts list per phase and price it properly, including the small items. Go through a catalogue for your model and actually list the seals, clips and fasteners. The total will be larger than you expect and that is the point.
Four: add the services — media blasting, machine work, plating, upholstery, glass, alignment.
Five: add consumables — abrasives, primer, sealer, filler, solvents, welding wire and gas, fasteners, fluids, tooling you do not yet own.
Six: add a contingency on the metal work specifically, because that is where the unknowns are concentrated.
Seven: compare the total to what a finished example sells for.
That last step is the uncomfortable one and it should be done anyway. On most vehicles the finished car is worth less than the cost of building it, and on common models it is worth considerably less. That does not mean do not build it. It means build it because you want that car, with your choices, knowing its history — and treat resale as a partial recovery rather than a return.
The people who get hurt financially are the ones who started with an investment thesis. The people who finish are the ones who knew the number before they started and decided it was worth it anyway.
Staging a Project You Can Actually Finish
If the total is more than you can commit at once — which is normal — stage it so that the car is in a stable, storable, reversible state at the end of each stage.
Stage one: make it solid. Structure and metal work, sealed and primed. At the end of this stage the car is protected and not deteriorating, which is the single most important thing. A primed shell in a dry space will wait for years without getting worse.
Stage two: make it stop and steer. Brakes, suspension, steering, wheels and tyres. Cheap relative to bodywork, and it makes the car movable, which is enormously useful.
Stage three: make it run. Drivetrain, fuel, cooling, exhaust, wiring. At the end of this you have a running, driving, unpainted car — and that is a genuine milestone that keeps projects alive, because you can drive it.
Stage four: shake it down. Find and fix every leak, noise, overheat and electrical fault while the car is still unpainted and you can get at everything. This stage is routinely skipped and it is why so many freshly painted cars get scratched and disassembled again chasing faults.
Stage five: paint. Now, with everything proven, with the car sorted, with nothing left to discover.
Stage six: interior and brightwork. Last, so nothing is climbed over or scratched.
Doing paint before the shakedown is the most common sequencing error in the hobby, and it costs real money every time. A fresh paint job on a car that has not been sorted gets damaged during the sorting.
Tooling, Space and the Costs Nobody Counts
A restoration has overheads that never appear in a parts list and that collectively amount to real money.
Space. The car occupies a bay for years, and so do the parts. A shell on stands, an engine on a stand, four doors in a rack, a hood, a trunk lid, four fenders, boxes of trim, a workbench and a parts washer need considerably more room than people plan for. If that means renting storage, it is a monthly cost multiplied by the length of the project, and on a five-year build it can rival a major phase. Worse, a project that gets moved into storage because the garage was needed is a project nobody works on.
Tooling you do not yet own. A welder and gas, angle grinders and a large supply of discs, a compressor with enough capacity for air tools and spraying, a spray gun, a media blast cabinet, an engine stand and hoist, a hydraulic press for bushings, a torque wrench in the right range, long sanding boards, bead rollers or a shrinker-stretcher for panel work, and a lift or at minimum good stands and a creeper. Individually reasonable, collectively substantial.
Compressed air is the one that surprises people. A compressor adequate for running an impact gun occasionally is not adequate for sustained blasting or for spraying paint, and discovering that halfway through is expensive.
Consumables. Abrasives are the big one and the invisible one. Block sanding a car consumes an extraordinary quantity of paper in several grades. Then welding wire and gas, cut-off and grinding discs, primer, filler, solvent, masking materials, gloves, rags, degreaser, seam sealer, cavity wax, thread locker, and penetrating oil by the gallon on a rusty car.
Services you cannot do yourself. Media blasting, machine work, plating, powder coating, glass, upholstery, alignment. Each has a going rate and none of them are negotiable if you want the result.
Shipping and freight. Panels, glass and complete assemblies ship as oversize freight, and on a project needing several large panels the freight alone is a line item. Returns on wrong parts cost freight twice.
Mistakes. Budget for them explicitly. You will buy a part that does not fit, cut a panel wrong, burn through a weld, or ruin a paint panel and have to redo it. Everyone does. A budget with no allowance for error treats every normal learning event as a financial crisis.
Your own time, valued honestly. Not because you should pay yourself, but because comparing "I will do it myself" against a shop quote is meaningless unless you count the hours. Four hundred hours at four hours a weekend is two years of every weekend, and that has a real cost in everything else you are not doing.
Paying a Shop: Getting It Right
If a shop is doing significant work, how the arrangement is set up matters as much as the rate.
Get the scope in writing. Not "restore the car" — a defined list of operations with a finish standard stated. Ambiguity here is where disputes come from, and the disputes are almost always honest misunderstandings about what "restored" meant.
Agree a phase at a time. A full restoration quoted as a single open-ended engagement is a bad arrangement for both sides. Metal work as one agreement, with a defined scope and an agreed hourly rate for discovered corrosion, is far healthier.
Agree what happens when more rust is found, because it will be. An hourly rate for additional work, with a call-back threshold before exceeding an agreed figure, protects everyone.
Ask for progress photographs. A shop doing good work is usually pleased to send them. A shop that goes quiet for months is a warning.
Understand the queue. Restoration shops run several cars at once and your car will sit for periods. That is normal. What is not normal is a car sitting outside, uncovered, partly disassembled.
Be clear about who buys the parts. Shop-supplied parts carry a markup and come with the shop's warranty and responsibility for fitment. Owner-supplied parts are cheaper and make fitment problems your problem. Both are reasonable; the misunderstanding is not.
Check that the shop will still exist. A multi-year engagement with a one-person operation carries real risk, and a documented, catalogued project is far easier to move elsewhere than a pile of unlabelled parts — which is one more argument for doing the teardown documentation properly even when somebody else is doing the work.
What Actually Keeps a Project Moving
A written specification, made before teardown, that you can hold yourself to.
One phase at a time, completely finished.
Documentation during teardown — photographs of everything, fasteners bagged and tagged by assembly rather than by size, and a written log of part numbers, torque values, clearances, suppliers and decisions. The log is what lets you resume after the inevitable six-month stall instead of starting the thinking over.
A contingency so that normal discoveries are not crises.
Visible progress. This matters more than it sounds. Finishing a subassembly — a complete rebuilt axle on a shelf, a finished door, a wired dash — sustains momentum in a way that a shell with four open phases does not.
A space you can leave it in. Many projects die because the garage was needed for something else and the car went into storage, and a car in storage is a car nobody works on.
The through-line in all of it is that the restorations that finish are not the ones with the biggest budgets. They are the ones where somebody worked out the real number first, decided it was worth it, and then refused to let the scope move.
Straight Answers
Common Questions
What is the single biggest cost in a restoration?
Metal work, bodywork and paint — consistently, and by a wide margin. It is almost entirely labour, it cannot be rushed, and the quality of the final finish depends on hundreds of hours of bodywork that nobody sees. Mechanical work, by contrast, is mostly parts with a predictable price.
How do I keep a restoration from spiralling?
Define the finish standard in writing before teardown — driver, show, or concours — and price each phase against that standard. Then do one phase at a time, fully finished, before starting the next. Projects spiral when several phases sit half-done and parts get lost, duplicated, or damaged in storage.
Will I get my money back?
On most vehicles, no. A full restoration usually costs more than the finished car is worth unless the model is genuinely rare and desirable. Restore because you want that vehicle. Treat any resale value as a partial recovery rather than a return on investment, and the maths stops being painful.
Is it cheaper to buy a finished car?
Almost always, and often dramatically so. A well-restored car typically sells for less than the cost of the restoration. Building one makes sense when you want specific choices, want to know every detail of its history, or want the experience itself — not when you want the cheapest route to owning one.