The 7-point operations leak audit you can run yourself

Most small teams do not have one big operations problem. They have seven small leaks, each quietly costing a few hours every week, that nobody has had time to name. This is the audit we run on the first call with every client, published in full. We implement these fixes for a living, we do not resell any of the tools named here, and our rate is the same whichever path you choose, so we have no reason to steer you anywhere except toward the leak that costs the most.

The research backs up the quiet-leak picture. Asana's Anatomy of Work Global Index found that knowledge workers spend 58% of their time on "work about work": chasing status, searching for information, switching between apps. Slack's Workforce Index in 2024 put it at about a third of the day lost to low-value work. Those are big-company studies, but the mechanics are identical at six people. The difference is nobody is watching the meter.

Before you start: the one number you need

Every cost formula below uses your loaded hourly cost: what an hour of a person actually costs the business, not their wage.

• Loaded hourly cost = (salary × 1.3) ÷ 2,000 hours.
• Example: a $75,000 salary is around $49 per hour. Call it $50.

The worked examples below use $50. Swap in your real number, with around 230 working days and 48 working weeks a year. The formulas are estimates, not invoices. Their job is to rank your leaks, so they run conservative on purpose.

Leak 1: No single source of truth

The real customer or donor data lives in someone's inbox, a personal spreadsheet, and two apps that disagree. Every decision starts with five minutes of figuring out which number is right. Gartner research put the cost of poor data quality at around $12.9 million a year for the average large organization. The mechanism scales straight down: people hunting for the right number, on the clock, every day.

Detect it in under 10 minutes

Ask two people the same question, separately: "How many active customers (or donors) do we have right now?" Compare the answers and where each person looked.

• Green (0): same number, same system
• Yellow (1): same ballpark, different places
• Red (2): different numbers, or "let me ask Sarah"

Size the annual cost

Formula: people affected × minutes per day spent hunting or reconciling ÷ 60 × loaded hourly cost × 230 days. Example: 4 people × 20 minutes × $50 × 230 days is around $15,000 a year. Twenty minutes a day is the gentle case. An older McKinsey estimate put time spent just searching for information at 1.8 hours a day.

The fix: DIY or hire it out

DIY: pick one system as the official record, your CRM or even one well-built spreadsheet. Announce a cutover date, move the side spreadsheets in, then lock or delete them. The hard part is the announcement, not the technology. Under 500 contacts, this is an afternoon and a decision, and you do not need us.

If we do it: consolidation and CRM cleanup is core RevOps work for us. Realistic range: 10 to 25 hours, so $1,500 to $3,750 at $150 per hour. The high end covers messy migrations across three or more systems with duplicate records.

Leak 2: Double entry

The same information is typed by hand into more than one place. A lead goes into the CRM and a spreadsheet. A payment lands in Stripe and gets re-keyed into accounting. In Zapier's survey of 1,000 knowledge workers at US companies under 250 people, 76% spent 1 to 3 hours a day just moving data from one place to another. A companion survey found data entry was the most dreaded task at work. Those surveys ran in 2021, and we still find this leak on most first calls in 2026.

Detect it in under 10 minutes

Ask one teammate to show you the last new customer, donor, or order they processed. Count the windows opened and the times they typed the same name.

• Green (0): typed once, everything else flowed automatically
• Yellow (1): typed twice for certain records
• Red (2): three or more times, or daily copy-paste is part of someone's job

Size the annual cost

Formula: records re-keyed per week × minutes per record ÷ 60 × loaded hourly cost × 48 weeks, plus 10% for fixing the typos. Example: 25 records a week at 6 minutes each is 2.5 hours. At $50 an hour over 48 weeks, that is $6,000, or around $6,600 with error cleanup.

The fix: DIY or hire it out

DIY: connect the two systems with Zapier or Make. The common pairs, web form to CRM, Stripe to accounting, ship as templates. Budget one afternoon and around $10 to $30 US a month for the tool. If both systems are mainstream and the fields map cleanly, you genuinely do not need a consultant.

If we do it: 4 to 12 hours per integration, so $600 to $1,800, under automation and AI builds. Hire help when one system has no public integration, when the field mapping is messy, or when an error in the flow would touch money or compliance.

Leak 3: Inconsistent follow-up

The work to earn each lead is already done and paid for. A slow reply throws it away at the last step. The classic MIT and InsideSales.com study, later popularized by Harvard Business Review, found that contacting a lead within 5 minutes instead of 30 made reps 21 times more likely to qualify it. Almost nobody clears that bar. When Workato tested 114 B2B companies with real demo requests in a study published in March 2026, over 99% missed the 5-minute window and the average email reply took almost 12 hours.

Detect it in under 10 minutes

Submit your own contact form from a personal email at 4 p.m. on a busy day. Time two things: the first acknowledgment, and the first useful human reply.

• Green (0): acknowledgment within minutes, human reply the same business day
• Yellow (1): a reply usually lands within a day or two
• Red (2): response time depends entirely on workload

Size the annual cost

Formula: monthly inquiries × share that go stale (no reply within one business day) × close rate × average first-year value × 12, then cut the result in half to stay honest. Example: 30 inquiries a month, 20% stale, 25% close rate, $2,000 average value gives $36,000 a year at face value. Call it around $18,000, since some stale leads come back anyway.

The fix: DIY or hire it out

DIY: set up an instant acknowledgment that includes a real booking link, plus a rule that flags any inquiry older than 4 business hours. Most form tools and CRMs include both on cheap or free plans. If you get fewer than 10 inquiries a month, skip automation entirely: a phone notification and discipline beat any build at that volume.

If we do it: a speed-to-lead build (routing, sequenced follow-up, tasks that nag until someone replies) runs 6 to 15 hours, so $900 to $2,250, inside our GTM engineering work. In the same Workato test, companies with lead routing averaged around 3.5 hours to first response, against around 13 hours without.

Leak 4: Reporting by scramble

If your board, your funder, or your own gut asked for this quarter's numbers today, how long would the answer take, and would you trust it? Reporting that hurts gets done late, wrong, or not at all.

Detect it in under 10 minutes

Open last quarter's report and try to reproduce one number in it, right now.

• Green (0): minutes, from a live dashboard you trust
• Yellow (1): a few hours of exporting and stitching
• Red (2): a multi-day scramble, every time

Size the annual cost

Formula: hours per reporting cycle × cycles per year × loaded hourly cost. Example: 6 hours a month at $50 is $3,600 a year. Then add what the formula cannot price: for a non-profit, a shaky funder report risks the next grant, usually worth more than every other leak combined. That risk is why funding-readiness work is its own service line for us.

The fix: DIY or hire it out

DIY: build one shared dashboard fed by the source of truth from leak 1. Google's Looker Studio is free, and most CRMs include dashboards on plans you already pay for. Start with five numbers, not thirty.

If we do it: 8 to 20 hours, so $1,200 to $3,000, depending on how many systems feed it. One warning: a dashboard on top of messy data is decoration. If you scored red on leak 1, fix that first.

Leak 5: The bus factor

If your most critical person left tomorrow, what breaks? Could anyone else run payroll, invoicing, or the funder report? Undocumented process is the leak leaders underestimate most, because the person carrying it rarely complains before resigning. It also blocks delegating, hiring, and real vacations.

Detect it in under 10 minutes

List your three most critical recurring processes. For each one, name a second person who could run it tomorrow using documentation that exists today.

• Green (0): all three covered
• Yellow (1): one gap
• Red (2): most of it lives in one person's head

Size the annual cost

This one is expected cost, like insurance math. Formula: chance the key person leaves this year (10 to 20% is realistic) × weeks to rebuild what they carry × their loaded weekly cost × 2, because the scramble pulls in everyone else. Example: 15% × 6 weeks × $2,000 × 2 is around $3,600 a year, per undocumented role. The formula understates it. It prices the rebuild, not the missed payroll run.

The fix: DIY, and only DIY

This is almost never a consulting job, and we charge $150 an hour to tell you so. Once a week, have the person record their screen while doing the task and narrating. Free screen recorders are fine. Drop the links into one shared document. After two months you have a runbook that cost an hour a week.

We map processes inside larger builds, 6 to 12 hours at $900 to $1,800, because a build needs a map. But paying outsiders to write down what your own people already know is slow and produces worse documentation. Do not buy it on its own, from us or anyone.

Leak 6: Tool spend on autopilot

Zylo's 2026 SaaS Management Index, built on data from 40 million licenses and $75 billion in spend, found the average organization leaves 36% of its paid SaaS licenses unused. It also found 78% of IT leaders hit surprise charges from AI or usage-based pricing. Those numbers skew enterprise, but seat creep looks the same at eight people: renewals on autopay, two tools doing one job, a seat for someone who left in March.

Detect it in under 10 minutes

Pull last month's card statement and list every software charge. For each, write down who owns it and what process it serves.

• Green (0): every tool has an owner, and the list was reviewed in the past year
• Yellow (1): a couple of question marks on the statement
• Red (2): honestly, no idea

Size the annual cost

Formula: total monthly software spend × 0.3 × 12. The 0.3 is a defensible default given Zylo's 36% unused-license figure. Example: $1,200 a month × 0.3 × 12 is around $4,300 a year.

The fix: DIY, and mostly DIY

Cancel, downgrade seats, and merge overlapping tools. This is one afternoon, twice a year. Put them in the calendar now.

Please do not hire us just for this. It takes you 3 hours and a credit card statement. We review tooling inside bigger engagements, 3 to 6 hours at $450 to $900, and the version that earns its fee is when one custom build replaces two subscriptions for good.

Leak 7: Dropped handoffs and the switching tax

When work crosses a team boundary, sales to delivery or intake to scheduling, does anything announce it, or does it travel by memory and hallway? Dropped handoffs are where customer experience quietly dies. They also force constant app switching. A Harvard Business Review study that tracked 137 workers at three Fortune 500 companies found people toggled between apps around 1,200 times a day and lost just under 4 hours a week to reorienting, about 9% of work time.

Detect it in under 10 minutes

Trace the last piece of work that crossed a team boundary. Ask both sides one question: how did you know it was your turn?

• Green (0): a trigger fired automatically, with a named owner
• Yellow (1): a checklist exists, and people mostly remember
• Red (2): someone found out when the customer asked

Size the annual cost

Formula: dropped or delayed handoffs per month × (rework hours × loaded hourly cost + revenue lost per incident) × 12. Example: 2 incidents a month, each costing 3 hours of rework at $50 plus $250 in lost or discounted work, is around $9,600 a year.

The fix: DIY or hire it out

DIY: a one-page handoff checklist, plus a single shared status field, like "ready for delivery", that sends a notification when it changes. Most project tools do this natively.

If we do it: status-triggered tasks, ownership rules, and an intake-to-delivery flow run 5 to 14 hours, so $750 to $2,100, usually as automation work or part of a RevOps cleanup.

Your score

Total points	Reading	What to do
0 to 4	Tight ship	Look for the one automation or AI step that turns good into effortless. Do not buy a rebuild you do not need.
5 to 9	Quietly leaking	Run the cost formulas on your yellows and reds. Fix the single biggest number, then stop and measure.
10 to 14	Costing you now	Resist the full-rebuild urge. Fix the source of truth or follow-up first. The other leaks usually share the same root.

The score is not a grade, it is a map. Its job is to make you fix the leak that costs the most, not the one that annoys you the most.

The one-page checklist

Print this table. Run one test a day for a week, or block 90 minutes and do all seven.

Leak	10-minute test	Annual cost formula	DIY fix	Our hours at $150
1. Source of truth	Ask 2 people the same data question, separately	People × min/day ÷ 60 × rate × 230	Declare one official system, migrate, lock the rest	10 to 25 hrs ($1,500 to $3,750)
2. Double entry	Watch one record get entered; count repeat typing	Records/week × min ÷ 60 × rate × 48, + 10%	Zapier or Make template between the two systems	4 to 12 hrs ($600 to $1,800)
3. Follow-up	Submit your own form; time the reply	Inquiries × stale % × close % × value × 12 ÷ 2	Auto-acknowledgment, booking link, 4-hour flag	6 to 15 hrs ($900 to $2,250)
4. Reporting	Reproduce one number from last quarter's report	Hours per cycle × cycles × rate	One free dashboard, five numbers max	8 to 20 hrs ($1,200 to $3,000)
5. Bus factor	Name a backup for your 3 key processes	Leave odds × rebuild weeks × weekly cost × 2	Weekly narrated screen recordings in one doc	Skip us. DIY this one
6. Tool spend	Owner and purpose for every charge on the statement	Monthly software spend × 0.3 × 12	Cancel, downgrade, merge, twice a year	Skip us, unless inside a bigger build
7. Handoffs	Ask both sides: how did you know it was your turn?	Incidents × (rework × rate + lost revenue) × 12	Status field, notification, one-page checklist	5 to 14 hrs ($750 to $2,100)

When you do not need a consultant

Under five people and 500 contacts, almost everything in this audit is a DIY job. If your team cannot agree on one source of truth, that is a leadership decision, and no consultant can make it for you. Leaks 5 and 6 are never worth outside money on their own.

Hire help, us or anyone, for the narrower cases: software that must talk to software with no template, flows where an error touches money or compliance, or reporting fed by several systems. Get the scope in writing first. Our rate is a flat $150 per hour, with no retainer minimums, and unused hours never expire. For a second opinion on your score, send it over. We will name the highest-value move on a 30-minute call whether or not you hire us to build it.

Frequently asked questions

Can I really run this audit without hiring anyone?

Yes. Every test uses things you already have: your card statement, your contact form, your last report, and ten minutes with a teammate. Several of the fixes are DIY too. Consultants only earn their fee on the integration-heavy leaks.

How accurate are the dollar formulas?

They are ranking tools, not invoices. Each one is built to be conservative, and the follow-up formula even tells you to cut the result in half. Use them to decide which leak to fix first, then re-run the formula 60 days after the fix to see what changed.

What should I fix first if everything scored red?

Source of truth or follow-up, in that order of preference. Most other leaks share the same root: data scattered across systems that do not talk. A full rebuild feels satisfying and usually stalls. One fixed leak, measured, builds the case for the next one.

What would it cost to have Tinyvation fix all of it?

The five fixes worth hiring out total 33 to 86 hours, so $4,950 to $12,900 at our flat $150 per hour. Almost nobody should buy that all at once. A focused first engagement is 5 to 20 hours aimed at the single worst leak, scoped in writing first. Unused hours never expire, so nothing pushes you to spend faster than the work demands.

Do these big-company statistics really apply to a team of six?

The mechanisms do, the magnitudes do not. Gartner's $12.9 million figure and Zylo's 36% unused-license figure come from large organizations, which is exactly why each leak here has a formula using your own numbers. Measure your leak. Trust your math over anyone's headline, including ours.