Connect with Software

Month: September 2026

The Contract Clause Test: How to Spot an SEO Agency Built to Be Fired Cleanly

The founder is on page 9 of a 14-page master services agreement at her kitchen table, coffee going cold, red pen in hand. The initial term is 24 months. Page 11 auto-renews it for another 12 unless she cancels inside a 30-day window that opens 90 days before renewal.

Page 12 says any content produced during the engagement stays the property of the agency until every invoice is paid, including invoices for months she hasn't reached yet. She hasn't made it to the reporting section. On the sales call, the agency described this contract as "pretty standard."

The best SEO agencies write contracts that make it easy to leave. That sounds backwards until you sit with it. An agency confident in its work doesn't need paper to hold you; it needs the results to.

The clauses in a services agreement tell you more about how a firm thinks about the client relationship than the pitch deck, the case studies, and the discovery call combined. The question isn't whether the contract protects the agency. It's whether it lets you walk away cleanly when the work stops earning its keep.

Decide How Long You're Willing to Be Locked In

Term length is the first real decision, and it reveals the most. SEO is genuinely slower than paid channels. 82% of surveyed experts put the point at which meaningful traffic starts to show at about six months, with full compounding closer to a year or two out. Agencies use that timeline to argue for long initial terms. The argument has some merit and a lot of self-interest.

A six-month initial term is defensible. It gives the technical audit, on-page work, and first content cycles time to register. Past that, the trade-off tilts against you. Month-to-month after an initial ramp is the cleanest structure, and if the agency won't offer it, ask what specifically breaks in month seven that requires you to be captive through month twenty-four.

Read the Renewal Clause Before You Read Anything Else

Auto-renewal is where reasonable-looking contracts turn into traps. Watch for a short cancellation window bolted onto a long renewal term: you have 30 days to opt out of another year, and that window happens to open 90 days before renewal, which you will not be tracking in month ten of feeling vaguely unhappy.

Regulators have noticed. The FTC's revised rules on recurring subscriptions reach most business-to-business auto-renew arrangements, not only consumer subscriptions. That's useful backdrop when you push back on a clause an agency insists is untouchable.

  • Match the notice to the term. A 30-day notice period against a month-to-month or quarterly renewal is fair. A 30-day notice period against a 12-month renewal is a trap dressed as a formality.
  • Kill silent renewal. Require the agency to send a written renewal offer that you affirmatively accept. If they can't be bothered to ask for the next year, they haven't earned it.

Decide Who Owns the Work When You Leave

Ownership language is where the friendly voice on the sales call goes quiet. The default in many agency templates is that deliverables (blog posts, landing page copy, schema markup, keyword research, even Looker Studio dashboards) remain the intellectual property of the agency, licensed to you for the duration of the engagement. The moment you leave, the license ends. You paid for content that walks out the door with them.

The clause you want is straightforward: all deliverables produced for you, and all data collected on your behalf, become your property on payment. Access to Google Search Console, Google Analytics, Ahrefs projects, and any CMS credentials the agency created must transfer back to you at termination, in a defined format, within a defined window.

Get the window in writing. "We'll send it over eventually" is how a botched handover turns into a quarter of lost lead volume you can't recover.

Decide What "Results" Actually Means on Paper

Every agency promises results in the pitch. Very few will define them in the contract. Force the decision: what specifically will be measured, how often, and what happens when the number doesn't move?

You don't need a ranking guarantee. Anyone offering one is either inexperienced or lying about how search works. What you do need is a written scope of work tied to leading indicators the agency can actually influence: pages published, technical fixes shipped, qualified backlinks earned, indexed URLs, non-brand organic sessions to money pages.

Google's own guidance on hiring an SEO suggests asking prospective providers how they measure success and whether they'll disclose every change they make to your site. Put both in the contract. If the agency resists that level of transparency in writing, they're telling you what the reports will look like once your card is on file.

None of this is hostile. It's the same energy a good agency brings to its own vendor relationships. For a short version of what to look for on the positive side, this rundown on the traits worth demanding from an the traits worth demanding from an SEO agency agency covers the ground well.

Bring that list to the sales call. Bring the red pen to the contract.

Perpetual vs Periodic Inventory for Marketplace Sellers

A periodic inventory system tells you what your cost of goods sold was after the quarter is over. A perpetual system tells you what it is right now. For a single channel seller with one warehouse and predictable turns, periodic is defensible and cheap. For a seller shipping from three fulfillment centers, a third party warehouse and a garage, periodic is a quarterly guess dressed up as a number, and every purchasing decision made between counts is made blind. That is the whole comparison. Everything below is about when the cheaper method stops being the rational one.

What the two methods do

Periodic inventory runs on one formula applied at period end. Beginning inventory plus purchases minus ending inventory equals cost of goods sold. You count what is left, and COGS is whatever the arithmetic says is missing. No entry is made when a unit sells. The general ledger carries a purchases account all period, and a single adjusting entry at close moves the balance where it belongs.

Perpetual inventory books the movement as it happens. Every sale reduces the inventory asset and recognizes COGS in the same transaction. The inventory balance in the ledger is supposed to equal the inventory on the shelf at any moment. Counts still happen, but they verify the system rather than produce the number.

The practical distinction is not accuracy at year end. Both methods land in roughly the same place once you have counted. The distinction is whether you have a usable number in week six.

The marketplace complication

Three things make marketplace selling harder on periodic than traditional retail.

First, your inventory is in places you cannot walk to. Units in an Amazon fulfillment center move between facilities without generating anything you would recognize as a transaction. Counting them means trusting a report from a party that is not your accountant.

Second, in transit inventory is real and often large. Goods on a container are your asset the moment title passes, which depends on the incoterms in your purchase order, not on when the pallet arrives. A periodic count taken while forty thousand dollars of stock sits on water will produce a COGS figure that is wrong in a specific and expensive direction.

Third, marketplace settlements net fees against revenue, which means the revenue side of your margin calculation is already obscured. Pairing an obscured revenue figure with a quarterly COGS estimate produces a gross margin that nobody should make decisions on.

A worked example

A seller starts the quarter with $180,000 of inventory at cost. During the quarter they buy $240,000. At quarter end they count $155,000 across all locations.

Periodic COGS is $180,000 plus $240,000 minus $155,000, which is $265,000. That figure is correct in aggregate and useless in detail. It does not tell you which SKUs consumed it. It does not separate shrinkage, damage and theft from actual sales, because all three land in the same residual. If $12,000 of that was a pallet damaged in receiving, periodic reports it as cost of goods sold, and your gross margin absorbs a loss that belongs somewhere else.

Perpetual would have booked $253,000 of COGS against specific sales and flagged a $12,000 inventory adjustment as its own line. Same total, completely different management information. One of those versions lets you find the problem in the receiving bay.

What the tax rules require

Less than most sellers assume, and the threshold is higher than most published guidance says.

Under section 471(c) of the Internal Revenue Code, a small business taxpayer meeting the gross receipts test may use a method that treats inventory as non incidental materials and supplies, or may conform to the method reflected in its applicable financial statement or books and records. The regulations were finalized in TD 9942. The inflation adjusted threshold is published annually: Revenue Procedure 2025-32 sets it at $32,000,000 of average annual gross receipts for tax years beginning in 2026.

One caution, because it trips people up. IRS Publication 538, the plain language guide most sellers reach for, carries a January 2022 revision date and still prints a $26 million threshold in more than one place. The number in the current revenue procedure governs. Check the revenue procedure, not the publication, and confirm with your own tax professional rather than treating any article as authority.

What this means in practice is that the choice between perpetual and periodic is usually a management accounting decision rather than a compliance one. The IRS will accept either from most sellers in this revenue range. Your bank, your buyer in a diligence process, and your own purchasing team are the parties who care.

When the switch pays for itself

Four signals, any one of which is usually enough.

You have stocked out of a top ten SKU in the last two quarters. Reorder points require knowing on hand quantity continuously, and periodic cannot supply it.

You sell the same SKU across more than one channel. Allocation between channels is invisible under periodic until the count.

Your gross margin swings more than a few points quarter to quarter without a pricing change. That volatility is usually a counting artifact.

You are preparing to raise money or sell. Diligence teams ask for monthly gross margin by product. Periodic cannot produce it retroactively.

The tooling question

Neither QuickBooks Online nor QuickBooks Desktop maintains marketplace inventory on its own, which is why sellers end up with a layer in between. The options are not interchangeable and each is better at something.

Webgility runs bi-directional inventory sync and maps to QuickBooks bins and locations, which is the deepest inventory feature set among the accounting focused tools and the right answer if warehouse operations are the constraint. Entriwise goes furthest on Amazon specific inventory events such as FBA inbound transfers and lost or damaged adjustments. Sellerboard maintains FIFO cost tracking and stockout adjusted forecasting, and it is better than most accounting tools at telling you when to reorder, though it is a seller analytics product rather than a bookkeeping one. ConnectBooks sits on the accounting side, syncing marketplace activity into QuickBooks Online, QuickBooks Desktop Enterprise and Xero with automated COGS and SKU level profit reporting. A2X deliberately does none of this and posts clean accrual journals instead, which is the correct choice for an accountant who wants the ledger right and the inventory system somewhere else.

Pick based on which problem is costing you money. A seller losing sales to stockouts and a seller failing an audit need different products, and the second one is cheaper to fix.

For the underlying accounting standards, the AICPA publishes guidance on inventory valuation, and the IRS small business and self employed pages cover the method change procedures. Changing an inventory method generally requires filing for a change in accounting method, which is not a decision to make in a spreadsheet the week before a filing deadline.

Inside the Control Room: The Mental Load of Running a High-Stakes Plant

At 3 a.m. on a night shift, a control room operator watches a wall of screens showing pressures, flow rates, and gas temperatures. Nothing is wrong. That's the point. For twelve hours, the job is to notice the one reading, out of thousands, that starts to drift, and to act before it becomes an incident.

The lights are low, the coffee is stale, and a bad decision in the next ten seconds can hurt somebody two states away.

This is the psychological reality of industrial work that outsiders rarely see. Plant operators, pipeline controllers, refinery techs, and offshore crews spend their careers in sustained vigilance, often on rotating shifts, often carrying the knowledge that a missed alarm can hurt people. The mental load doesn't show up in a safety report. It shows up in sleep, in relationships, in the slow erosion of the person doing the watching.

Deciding Whether the Job Is the Problem, or the Schedule Is

Operators who feel worn down tend to blame themselves before they look at the roster. That's usually the wrong order. Shift design is one of the biggest drivers of fatigue-related risk in industrial work, and the evidence on it is not subtle. OSHA notes that accident and injury rates climb sharply on evening and night shifts, and that long shifts of twelve hours or more raise injury risk considerably, the same conditions many plants normalize as standard rotation.

Before you decide you're burned out on the work itself, look honestly at the pattern you're being asked to sustain. Two useful questions:

  • How many consecutive nights? Rotating patterns that flip direction every few days give the body no time to adjust and read to your nervous system as chronic jet lag.
  • How real are the breaks? A break where you're still on radio and still watching the board is not a break. If nobody physically takes the seat, the vigilance never stops.

If the schedule is the problem, resilience training alone is unlikely to fix it. Have that conversation with the plant, not with yourself.

Name Vigilance Fatigue Before It Names You

Vigilance fatigue is a specific thing, and it doesn't feel like being tired. It feels like the screen going flat. Alarms register a half-second late. You reread the same trend line three times.

Federal pipeline regulators have long acknowledged that control-room monitoring is a sedentary, high-vigilance task where fatigue is a predictable hazard rather than a personal failing, a point made explicit in PHMSA guidance on control room fatigue countermeasures.

The decision here is whether to treat that dulling as normal. Don't. When your attention flattens on shift, that's data. Log it the way you'd log a pressure anomaly.

Patterns matter more than any single bad night, and a written record gives you something concrete to raise with a supervisor or an occupational health contact, instead of a vague sense that something is off.

Deciding What the Plant Owes You, and What You Own

There's a split of responsibility worth being clear about. The plant owns the conditions: staffing levels, shift rotation, alarm design, control room ergonomics, and whether the equipment itself is engineered to reduce operator load. Well-designed process skids, the kind that arrive pre-engineered, pre-tested, and instrumented for remote monitoring, like the fuel gas conditioning systems used across power generation and upstream oil and gas, take real cognitive load off the operator by making the process legible and predictable. Bad engineering pushes that load back onto humans, and humans absorb it in cortisol.

What you own is narrower but real: sleep hygiene between shifts, honest conversations with the people you live with, and the willingness to use whatever support the employer already offers before you need it in a crisis. Employee assistance programs, peer-support networks, and structured debriefs after serious incidents exist for a reason. Using them early is cheaper, in every sense, than using them late.

Ask for Help Early, and From the Right Person

The hardest decision in this whole picture is the personal one. Industrial workforces skew toward stoicism, and stoicism has a cost. If a specific incident has stayed with you, a near-miss, a coworker's injury, an alarm that ended badly, a structured critical incident debrief within days of the event is far more effective than trying to walk it off for months. If the weight is diffuse and long-running, that's a case for a clinician who understands shift work and occupational stress, rather than a general wellness app.

None of this is soft. Attention is the safety system, judgment is the safety system, and the person watching the board is the safety system.

Treating that person's mental health as core operational hygiene, rather than an HR afterthought, is how high-stakes plants stay high-stakes plants instead of case studies.

© 2026 PSPL

Theme by Anders Noren — Up ↑