Every estimator has had the bid that got away. You priced it right, you knew the project, and the GC came back with feedback that someone else hit the same scope two days faster. Or worse: you won the work, the steel hit the shop, and the BOM was off because a detail was missed on a sheet nobody reviewed twice. The first kind of loss is competitive. The second kind costs you margin.
Both problems trace back to the same root cause: a lack of systematic discipline across the estimating workflow. The best steel estimators do not just count beams faster than everyone else. They follow consistent practices that compound across hundreds of bids per year, turning individual expertise into team-level capacity. The shops that systematize these practices bid more work, win more profitable jobs, and grow without proportional headcount.
This article walks through ten practices that separate high-performing steel estimators from those running on instinct alone. It is meant for estimators with five or more years of experience who already know the basics and want to systematize what they do. It is also meant for estimating managers and shop owners who need a framework to evaluate whether their team's process actually scales.
This article sits under Building a High-Performance Steel Estimating Workflow and covers the practices that, applied consistently, turn that workflow into competitive advantage.
The most expensive estimates are the ones that should never have been bid in the first place.
Bid-no-bid decisions are where the highest-leverage time savings live in any estimating department. A bad-fit project consumes 40-80 hours of estimator time, generates no revenue, and crowds out a better-fit project that could have been won. Yet most shops make these decisions on gut feel without a structured qualification step.
A rigorous qualification framework asks five questions before any takeoff hours go in:
Maccabee's Don Fleszar captured the math behind why bid volume matters: "If we can increase the number of bids we put out by 50 to 100 percent, we're going to increase the amount of work we have equivalently." But that math only works if you're bidding the right projects, not just more of them. Discipline at the qualification step is what makes capacity gains profitable rather than exhausting.
A shop with $5-50 million in annual revenue typically can pursue 100-300 estimates per year. The qualification step decides which of the available opportunities deserve those hours. The estimators who track this systematically (win rate by project type, by client, by geography) build a feedback loop that improves the next year's qualification decisions.
The mistake to catch is the missed scope at hour 40 of takeoff, not the missed scope at hour 4.
Most steel estimators have learned the hard way that the time invested in a structured plan review before takeoff starts is the cheapest hour in the entire process. The drawing review is where you find the gaps, conflicts, and ambiguities that will otherwise drive RFI cycles, scope disputes, or worse, change orders against your own margin after award.
A thorough plan review covers:
The cost of skipping this step is real. According to the Construction Industry Institute, rework represents between 2% and 20% of total project costs, with an average of 12%. Not all rework traces to estimating gaps, but takeoff sits at the front of the chain, and a missed detail at the bid stage compounds through procurement, fabrication, and field installation. The plan review is the cheapest place in the chain to catch it.
For a deeper checklist on what to verify, see The Precision Gap: Why "Automated" Takeoff Software Is Failing Steel Estimators.
When two estimators in the same shop produce different BOMs from the same drawing set, you have a workflow problem, not a knowledge problem.
Consistency is what makes estimating institutional rather than personal. The estimator who has been doing this for fifteen years has built up a personal methodology that works for them, but if that methodology cannot be documented, taught, and reproduced, the shop is one resignation away from a capacity crisis.
The practices that drive consistency:
This is exactly where AI takeoff tools like LIFT add value without replacing estimator judgment. The AI applies the same categorization logic on every project automatically. The estimator owns the methodology choices (what counts as a connection, what attributes matter for this project), but the execution stays consistent regardless of who is running the takeoff that day. For more on what the AI actually does on a drawing, see How AI Reads Structural Steel Drawings.
A takeoff is only as good as the prices applied to it.
The shops that consistently win profitable work do not pull material prices from memory or last quarter's purchase orders. They maintain living databases that reflect current market reality. Steel prices move with commodity cycles, mill availability, and freight markets in ways that can erase a margin overnight if your database is stale.
The components of a useful material database:
The discipline is not just keeping the database current but reviewing it after each won project. Did the actual purchase price match the estimated price? If not, why? Was it a market move you couldn't have predicted, or a database stale-data problem? The estimators who close this feedback loop end up with material pricing that is consistently within 1-2% of actual, which is the foundation for competitive bidding without margin erosion.
The single biggest variable in steel pricing accuracy is labor hours, and labor hours are where most estimators are flying blind.
Material is relatively easy to price because the inputs are public (mill prices, freight, coatings). Labor hours are private to your shop and depend on equipment, workforce skill, project mix, and the specific connection complexity of the job in front of you. Shops that win consistently track their actual production rates against their estimated rates and update their assumptions accordingly.
What to track:
The ratio that matters: estimated hours to actual hours. Shops that consistently land within 5-10% on this ratio have a real estimating discipline. Shops where the variance is 25-40% are essentially guessing with seniority backing the guess, which works until the labor market or the project mix shifts.
Tracking this requires post-project actuals, which connects to best practice #9 below. The estimators who treat post-project review as a mandatory step, not an afterthought, are the ones whose rates stay accurate over years.
Markup is where strategy lives, but most shops apply markup by rule of thumb rather than by deliberate framework.
A systematic markup framework considers four inputs:
The CFMA Construction Financial Benchmarks Report shows industry net profit margins running around 5-6%, with specialty trades around 6.9% and heavy industrial around 4.1%. At those margins, even a small markup error compounds quickly. A 2-percentage-point pricing miss on a $2 million project is $40,000, which is most of the net margin on a separate job.
The shops that maintain consistent margin do not just apply a markup percentage. They apply it deliberately, based on each project's specific risk and strategic profile, and they document why each bid carried the margin it did so the framework gets sharper over time.
Curious whether your team is ready to systematize these practices? 5 Signs Your Steel Estimating Process Is Ready for an AI Transformation is a quick gut-check on where your shop sits today.
The bid that goes out without a second set of eyes is the bid that will eventually cost you significant money.
Multi-level review is the discipline that separates shops winning consistently from shops winning by luck. The principle is simple: nobody catches their own mistakes well, especially under time pressure, and especially after spending 30+ hours immersed in the same drawing set. A structured review process catches the errors that the original estimator literally cannot see anymore.
A practical three-level review for serious bids:
This process feels slow, especially on tight bid timelines, but it is the highest-ROI discipline in estimating. The math is simple: catching a 1% pricing error on a $1 million bid is $10,000. The peer-review hour that caught it is $50-100 of estimator time. The ROI on disciplined review is several orders of magnitude per hour invested.
This is also where AI tools enable better review, not worse. When the BOM is auto-generated and traceable back to the drawing in one click, peer review becomes faster and more focused, not skipped because it's tedious. Maccabee estimator Dawn Hargraves described the dynamic directly:
"I actually appreciate that it's not 100% perfect because it keeps me engaged and checking the work. We can catch any issues while still saving massive amounts of time."
That is exactly the partnership model the human-in-the-loop research literature recommends: AI provides the consistent baseline, the estimator's review is the discipline that catches what the AI missed.
Checklists are the most underrated tool in steel estimating.
The reason checklists work is that they convert tacit knowledge (the things experienced estimators just know to check) into explicit knowledge (the things every estimator in the shop will check). They are the lowest-cost, highest-leverage way to standardize quality across a team.
Two checklists every shop should maintain:
Pre-takeoff checklist (10-12 items):
Pre-submission checklist (8-10 items):
Checklists are not bureaucratic overhead. They are how shops maintain consistent quality across estimators with different experience levels. For more on the broader QA discipline, see AI Errors and How to Catch Them: Quality Control Best Practices.
The estimators who get better over time are the ones who close the feedback loop between what they estimated and what actually happened.
Post-project review is where institutional knowledge gets built. Every won project is a data point that should refine the next bid's assumptions. The shops that skip this step are essentially running on the same assumptions year after year, watching their accuracy drift as the market shifts around them.
A useful post-project review covers:
This data, accumulated over 50-100 projects, becomes the foundation for everything else: better production rates, better markup decisions, better project qualification. It is the difference between an estimator who has been doing the job for 15 years and an estimating department that has been getting better for 15 years.
The discipline is to make this mandatory, not optional. The shops that win consistently treat post-project review as a standing process, with a recurring meeting and a structured template, not an after-the-fact exercise that happens when someone remembers.
The senior estimator who retires takes 20 years of judgment with them unless that judgment has been captured systematically.
Estimating logs and reusable templates are how shops protect themselves against the inevitable estimator turnover that the labor market will keep delivering. Construction Dive's analysis of the estimator talent gap cites AGC data showing one in four construction workers is over 55 and the BLS projects 41% of the current workforce could retire by 2031. The BLS Occupational Outlook Handbook projects cost estimator employment to decline 4% from 2024 to 2034. The estimators you need are aging out faster than they are being replaced.
What to capture in your estimating knowledge base:
This is the work that turns individual expertise into team capability. Without it, every senior estimator's departure is a capacity crisis. With it, the shop's competence compounds over time even as people come and go.
Reading through these ten practices, the pattern should be clear: they all require systematic discipline applied consistently across hundreds of bids per year. That is exactly the kind of work where AI augmentation provides the most leverage. The Stanford Digital Economy Lab's research on AI augmentation vs automation makes the point empirically: when AI use is augmentative (supporting the estimator's judgment), employment in that occupation actually grows.
AI takeoff tools like LIFT support these practices in specific ways:
For more on the broader capacity implications, see How AI Multiplies Estimator Capacity (With Real Examples) and Breaking the Headcount Barrier: Scaling Bids Without Hiring.
The customer evidence backs this up:
The pattern across all five shops is the same: AI did not replace estimator judgment. It amplified the practices the estimators were already trying to apply.
These ten practices are not new. Good estimators have been doing some version of them for decades. What is new is the pressure on estimating departments to scale: more bids, faster turnaround, leaner teams, against a labor market that is shrinking. The shops that win the next decade will be the ones that systematize these practices and use AI to amplify them, not the ones running on senior estimator intuition alone.
The starting point is honest self-assessment. Pick the practice on this list where your shop is weakest. Pre-bid qualification? Plan reviews? Material database maintenance? Production rate tracking? Make that the first one you systematize over the next 90 days. Then move to the next.
If you want to see how AI tools support these practices in production, the simplest test is to run an upcoming bid through LIFT in parallel with your current process. Compare both the time and the quality of the output, and decide where AI fits in your workflow. You can start by booking a live demo.
