PE Firms Look at 4 Numbers, Not 40. Here's What They Actually Evaluate.

I spent years in private equity evaluating contractor roll-ups. Owners often showed up with elaborate dashboards. Most of those rows rarely drove the initial decision.

These four operating checks frame an initial diligence screen alongside verified EBITDA and revenue quality. They are not an exhaustive list, and no metric determines a fixed multiple. If you're 12-24 months out from an exit and trying to maximize valuation, focus here.

The 4 numbers that move PE valuations

1. DSO

Days Sales Outstanding. How fast you collect cash after invoicing.

Why PE cares: Lower DSO means less cash tied up in receivables to run the same revenue. At the same revenue, 90-day DSO carries 50% more receivables than 60-day DSO (90 ÷ 60 = 1.5). DSO reaches price mainly through the purchase agreement. Buyers usually set a normalized net working capital target (the peg) from trailing months, and the price adjusts up or down if delivered working capital differs at close. A sudden DSO improvement right before close can produce a cash release and a peg negotiation at the same time, so show the trend over 12 months.

The math (fictional, with stated assumptions): Assume $50M of annual revenue, $5M of EBITDA and an 8% cost of capital.

• At 90-day DSO, receivables are about $12.33M ($50M × 90 ÷ 365), and carrying them costs about $986K a year.
• At 45-day DSO, receivables are about $6.16M, costing about $493K a year.

These are two different effects. Moving from 90 to 45 days releases about $6.16M of cash once, as a balance change, and lowers annual financing cost by about $493K. Neither is recurring EBITDA, and neither becomes enterprise value by multiplying it by a deal multiple.

Counterexample: suppose the peg is set from a trailing average that reflects 90-day DSO, and the seller delivers 45-day AR at close. In a cash-free, debt-free deal the seller keeps the collected cash but delivers working capital below the peg, and the price adjusts down by a similar amount. The durable benefit is the lower working capital need going forward, which shows up in the buyer's cash-flow model.

Valuation impact: This page cites no transaction sample showing a DSO multiple effect, and the Level benchmarks cited here do not include DSO. Treat any multiple change as the buyer's judgment. For scale only, a 0.5 to 1.0 turn difference on $5M of EBITDA would be $2.5M to $5M of enterprise value. That is a fictional sensitivity, not an expected result.

Where most contractors stand: The nearest Level benchmark is collection rate, not DSO. Median contractor collection rate is 85.1%, per Level Index published metric (see its eligible sample and definition) across 464 contractors, defined as cash collected as a share of billed revenue.

P75 is 92.7% and P90 is 96.0%, separately labeled comparison points. That band is not a DSO target or a buyer requirement. The gap between your rate and 92% is also not all recoverable cash, because part of it may be disputed, retained or uncollectible.

Measure DSO separately: ending AR ÷ trailing 12-month revenue × 365. Billing speed matters too. In the billing-speed sample (n=733 total), the raw median is 1 day including progress billing. The post-completion subset has a 7-day median; its separate count is not published. DSO is a balance-to-revenue measure. Track completion-to-invoice and invoice-to-payment timing separately.

2. Gross margin per job type

PE doesn't care about your blended gross margin. They care about gross margin by service line: install, service, maintenance, T&M.

Why PE cares: Buyers want to know which revenue is recurring and which margins hold up. This page has no measured install-versus-service margin ratio, so do not assume service earns a fixed multiple of install margin. Measure your own. At the same EBITDA, a buyer may still prefer a mix with more recurring agreement revenue because it is easier to forecast, but that is a judgment about durability, not a formula.

The math: Level has two measured reference points, each with its own population.

• Job-level gross margin: across 1,747,089 completed jobs from 1,791 companies, the median job-level gross margin is 44.3%, per Level Index published metric (see its eligible sample and definition) across all job types (p25 28.4%, p75 60.2%).
• Service-agreement gross margin: across 259 companies with at least $10K of service-agreement revenue, the median is 37.9% (p25 20.5%, p75 53.5%).

Neither is split into install, T&M or service-call bands, so build your own table:

Operator checks:

• Use one consistent labor cost basis per job. If loaded technician cost is already in job cost, do not also count it again as overhead to be recovered.
• If your estimates use markup, convert to margin before comparing. A 50% markup on cost is a 33.3% gross margin.

Valuation impact: This page cites no transaction source for a multiple band by service mix, so none is stated as fact. For scale only: on $5M of EBITDA, each one-turn difference in multiple is $5M of enterprise value. That is fictional sensitivity arithmetic, not a market multiple. Keep the earnings measure consistent too. A multiple quoted on EBITDA cannot be applied to SDE or to a discounted cash-flow value without restating the base.

The catch: Many contractors do not have job-cost data clean enough to break this out. If you cannot show service versus install margin with labor, material, and subcontractor costs attached to each job, a buyer may apply conservative assumptions, widen diligence adjustments or move value into deal structure.

3. Customer concentration

What percentage of your revenue comes from your top 1, top 5, top 10 customers.

Why PE cares: Concentration is single-customer risk. If 30% of revenue comes from one customer and that customer leaves, revenue falls 30%. Profit can fall further, because overhead does not leave with the customer.

Fictional example: $10M revenue, $1M EBITDA, a 30% customer and a 35% contribution margin on that customer's work. Losing $3M of revenue removes about $1.05M of contribution, more than the whole $1M of EBITDA, unless costs are cut. Deals that use acquisition debt also face lender review of that risk, under each lender's own credit policy.

The math: This page cites no sourced lender or buyer concentration threshold, so it does not state one. Across 959 companies with more than $100K of annual revenue, the median company's largest customer accounted for 31.0% of total revenue (p25 18.5%, p75 54.6%, p90 82.4%). That is a cross-trade benchmark, not a buyer cutoff.

To measure your own:

• Calculate top 1, top 5 and top 10 shares on the same trailing 12 months and the same revenue definition.
• Check whether the largest relationship is one contract or several sites under one parent.

Valuation impact: No sourced discount schedule is cited here. Buyers can respond to high concentration through price, structure or both. Examples include an earnout, a holdback, or a requirement that the key customer contract can be assigned or survive the change of control. Read the contract's assignment and termination clauses before the buyer does.

The fix: As a planning assumption, allow 12-24 months, because new customers need trailing revenue history before a buyer counts them. Diversify deliberately, but check the margin. Winning smaller customers at a negative contribution margin lowers concentration and EBITDA at the same time.

Counterexample: adding $1M of new revenue at a 5% contribution margin cuts the top-customer share from 30% to about 27% ($3M ÷ $11M) and adds only $50K of contribution.

4. Estimated vs actual labor variance

How accurately you estimate labor hours on jobs vs. how many hours actually get logged.

Why PE cares: This is a useful proxy for operational discipline. As an illustrative diagnostic, labor-hour variance within +/- 5% suggests estimated and actual hours stay close. It does not by itself prove correct pricing or on-time completion. If variance is +/- 25%, the contractor is essentially gambling on every job.

The math: Across 315,393 jobs with both budgeted and actual labor hours, per Level Index published metric (see its eligible sample and definition), the median job lands at 99.4% of budgeted hours, on target. The mean lands at 119%, because 40% of jobs go over their hour budget and 18.3% run past 150% of it. A buyer who only sees the median thinks estimating is tight. A buyer who runs the distribution finds the tail.

Valuation impact: This isn't a direct valuation lever. It shapes how much a buyer trusts the job-cost data behind the other numbers. No sourced multiple premium is stated here. Tight estimating can shorten diligence questions and reduce the adjustments a QoE review proposes.

The fix: Easier to fix than concentration but harder than DSO. Requires field-level discipline: every job estimated against historical data, weekly variance review, root-cause analysis on outliers. As a planning assumption, allow 6-12 months to clean up.

What buyers actually do in due diligence

This is an illustrative diligence sequence, informed by the author's public private-equity background. It is not a verified engagement count or promised timetable; timing varies by buyer and deal.

Week 1: Initial outreach. Buyer wants 3 years of P&Ls, top customer list, employee headcount, basic ops description.

Week 2-3: Buyer reviews materials. Builds a quick valuation model. Decides if there's enough meat to justify a real bid.

Week 4-6: First bid (LOI) submitted. It is often a range rather than a specific number, conditioned on quality of earnings. The range is that buyer's, not a market standard. The 4 numbers above, alongside verified EBITDA and revenue quality, can influence where you land in that range.

Week 7-10: Quality of earnings (QoE) review. Independent accountants come in and verify EBITDA, normalize for one-time items, dig into job-cost data. This is where your concentration, DSO, and gross-margin-by-job-type get scrutinized in detail. It is also where the normalized net working capital target (the peg) is usually proposed. Expect the QoE to test whether job-level labor, material and subcontractor costs tie to the general ledger.

Week 11-14: Final bid. Specific. Lower than the initial range if QoE found problems; held at or near the range if everything checked out clean.

Week 15-26: Negotiation, contract, closing.

Total timeline: 6-12 months from first conversation to close. But the four numbers are determined long before week 1.

The 18-month exit prep that moves the 4 numbers

If you're 18 months out and want to maximize multiple:

Months 1-6: Audit and triage.

• Pull last 12 months and calculate the 4 numbers honestly. Don't game them.
• Identify the gaps against your own trailing history and the benchmarks above, keeping each benchmark's population in view.
• Pick the 1-2 that move most for your specific business.

Months 7-12: Operational fixes.

• DSO: tighten collection cadence and model your billing speed. Before using a day-30 escalation step, check each contract and state rule for lien notice, prompt-pay, retainage and dispute deadlines. Those legal dates override an aging playbook and can fall earlier.
• Margin by job type: clean up job-cost data, reprice low-margin work, drop loss-making customers.
• Concentration: deliberate sales diversification, smaller-deal win rate increase.
• Labor variance: weekly variance review, estimating retraining, post-mortem on outliers.

Months 13-18: Stabilize and document.

• Make sure improvements show in the trailing 12 months of financials before you go to market.
• Build the books PE wants to see (clean QuickBooks, monthly close in under 10 days, no big adjustments).
• Engage a sell-side advisor at month 16 to start running the process.

What NOT to bother fixing

Things contractors stress about that PE buyers don't care about:

• Marketing CAC. PE doesn't run marketing the way you do, they're going to roll you up with other contractors and rationalize spend.
• Vehicle fleet age, as a cosmetic issue. Buyers model fleet replacement as capital spending that reduces free cash flow, so have a replacement schedule and maintenance records ready. Buying new trucks right before a sale mainly converts cash into assets the buyer was already modeling.
• Office space. Lease assumption or termination is often a contract detail; related-party or off-market rent can require an EBITDA normalization and affect value.
• Software migration. Don't switch tools right before a sale, because a buyer may consolidate systems after close. Do make sure the current stack can export job-level labor, material and subcontractor costs that tie to the general ledger. The QoE review depends on that data.
• Employee retention plans for the office staff. Field-team retention matters, and buyers may also focus on key finance, estimating and dispatch roles.
• Branding / reputation polish. They're buying cash flow, not your logo.

If you spend the 18-month exit prep window on these, you're polishing things that don't move price. Spend it on the 4 metrics above.

The thing nobody tells you

PE buyers often know more about how businesses like yours get diligenced than you do.

Experienced buyers have often reviewed other contractors in your trade and bring their own comparison points. Assume they will rebuild your numbers from source data: AR aging, job-cost exports, customer revenue by month and the general ledger.

Trying to spin numbers doesn't work. They'll find it in QoE and re-trade you down. The honest play is: clean numbers, real improvements, durable business. The owners who exit at the high end of multiples didn't pull a magic trick at month 23. They built durable financial discipline over the prior 5-10 years and the four metrics happened to be in the green when buyers came calling.

If you're 18 months out and one of the 4 numbers is materially red, fix that one. Forget the dashboard. Start with these four checks, then reconcile EBITDA, revenue quality and the rest of the diligence evidence.

Find your gap in 2 minutes or book a free 30-min audit to walk through the 4 numbers using your own AR aging, job-cost and customer revenue data.

The multiple moves too

Some operating improvements can raise EBITDA, while shorter DSO principally releases working capital. A buyer may price risk differently, but neither a multiple uplift nor an EBITDA increase is automatic. The band arithmetic is in Why 25% More EBITDA Can Be Worth 88% More Company.

Related reading

• How PE firms evaluate contractor financials
• Exit readiness for contractors
• Public contractor valuation multiples
• Customer revenue concentration benchmarks
• Maintenance vs install revenue (and why valuation differs)