Ivan

One social trick that works better than any tooling: reply with your best guess at what went wrong and ask them to confirm. People who will not answer a list of questions will absolutely tell you that you are wrong, in detail, with steps. It feels backwards and it halves the round trips on vague tickets.

That framing helps - I was treating it as one tax problem instead of two payrolls that can't see each other. Adjusting the smaller one makes sense because the salary there is easier to absorb.

Before you consolidate, cull. I did the same audit with 31 domains and the useful exercise was a spreadsheet with four columns: domain, renewal date, renewal price, and one sentence saying what it is for. Any row where I could not write the sentence got dropped.

Nine went. Nothing bad happened. Two of them I had been renewing for six years for an idea I could no longer remember. The consolidation question matters much less once the list is short enough to hold in your head, and the annual cost fell by more than half before I moved anything.

I did the version where you find out afterwards. Two small contracts over about eight months, maybe four thousand dollars total, on the basis that a colleague said it was fine. When my paperwork came up later I ended up disclosing it and paying for several hours of attorney time to explain it properly, which cost more than the contracts had paid and took two months of sleeping badly. It resolved. I would pay for the consult first every single time now.

The clause I use is a background IP carve-out and it's two paragraphs. In plain terms: anything I bring to the engagement, plus anything general-purpose I write during it, stays mine; the client gets a perpetual licence to use it as part of their deliverable; the client owns the parts specific to them: their branding, their business rules, their data, their integrations.

A solicitor drafted mine and it cost a few hundred, once, and I've used it on every contract since. Mine, not yours: the enforceability of that shape varies by jurisdiction and by how your contracts are structured, so take it to someone qualified before you rely on it.

The other half is structural: separate repositories from day one. A carve-out is much easier to argue when the shared core is a package with its own history and the client repo imports it.

Then the new-contract answer matters more than the old-contract answer, which is what you asked. Start the clean package now, build the next engagement on top of it, and let the history start being useful from today.

Going to disagree with the people telling you to wing it. For a first trip to a continent where you don't speak the language, the tour buys you a floor under the worst version of the trip - the day you're ill at altitude, the day a road is closed, the day you get the bus time wrong. Ten days with strangers is not a marriage, and every tour I've been on had people peeling off for the afternoon without drama. Do it independently on trip two, when you know what you actually like doing.

Practical warning from doing this across two banks: the shared-app setups that depend on one person's login are fragile. Ours was linked under my account, and when I changed banks the whole joint history broke and my partner couldn't fix anything because none of it was really hers.

If you can, connect the joint account under whichever of you is least likely to change banks, and export a CSV backup quarterly. Ten minutes a quarter against losing three years of shared history.

It doesn't exist because it's a support nightmare to build. Choosing which accounts to connect is the permission system, and it already works.

That is the expected behaviour and it catches a lot of people. The residential clean energy credit is nonrefundable, which in plain terms means it can reduce the tax you owe to zero but it cannot pay you money you never owed. The IRS page for it is explicit that you can carry forward any excess unused credit and apply it against tax you owe in future years.

So the number you should have been looking at when you were doing the back of the envelope was your total tax liability for the year, not the refund you normally receive. I am not your accountant and if the carryforward is a big number it is worth an hour with a preparer, but nothing about what your software did sounds wrong.

If you genuinely can't get either to bend, that's data rather than an obstacle. A role that requires your live presence at a fixed time every single morning, no exceptions, is a role that is structurally hard to hold alongside anything else. I'd start thinking about which of the two is the one you keep.

The argument against add-ons that nobody makes until they have lived it: every SKU you add multiplies your billing surface. Mid-cycle additions produce prorations, cancellations produce credits, annual plans with a monthly add-on produce alignment questions, and every one of those becomes a support ticket that you answer personally.

I have two SKUs and they generate more billing confusion than the product generates product confusion. One price is a feature.

Most apps do show both if you tap into the balance. It's just that the big friendly number on the front screen is nearly always the flattering one.

It's waiting to find out whether the money is real. Two separate things are happening and your app is showing you both in the same column, which is why it's confusing.

The number you can see is the ledger balance, which is what has been recorded against your account. The number you can spend is the available balance, which is what the bank is willing to let go of before it's certain the funds have actually settled.

A cheque in particular can come back unpaid days after it was deposited. If the bank released the cash immediately and the cheque bounced, they'd be chasing you for money you'd already spent. The hold is them absorbing that risk on a schedule instead of case by case.

Electronic transfers depend entirely on the system used. Some run in batches a couple of times a day and settle the next business day, some are genuinely instant and irrevocable. Weekends and holidays don't count as business days, which is how a Friday deposit turns into Tuesday.

How long any of this takes varies a lot by country and by bank, and your account terms will spell out the actual availability rules if you want the specific numbers.

For the business side of this: I have one paragraph in a support document that says target first response next business day, business days defined, no guarantee of resolution time, and a note that critical outages are handled outside those hours on a best effort basis. That has passed several procurement reviews without anyone blinking. What kills deals is having nothing written down, not having a modest number written down.

True for self serve, false the moment you sell to a company with a procurement process. The first thing their form asks for is a response time commitment, and no number on the site means an awkward email at exactly the wrong stage of the deal.

It reads as prioritisation, not weakness, as long as you bring the list and not the feelings. Managers who've just cut a team are used to this conversation and often relieved someone started it.

The workload at A is not going back down. Post-layoff redistribution almost never reverses, because from management's side the experiment succeeded, the work still got done with fewer people.

So the question isn't how to survive this quarter, it's which of the two you're keeping, and you should decide that on purpose rather than by finding out which one fires you first.

In the meantime, make the load visible at A. Written list of everything you now own, sent to your manager, with a line asking them to rank the top five because you can't do all of it well. That does two things: it either gets some work moved, or it gets you a paper trail showing you flagged it. Both are useful.

Half agree. The maintenance is real, but so is the moment in year three when the client wants something WordPress does out of the box and your static setup needs a rebuild. I price the maintenance in rather than avoiding the platform.

The trick that solved this for me was stopping the client from buying hosting at all. I quote a care plan that includes hosting, backups, updates and a fixed number of small changes, and I buy the hosting wholesale under my own account. One line item, one number, no renewal shock, and I control the environment which means I control how long a problem takes me to fix.

The cheap shared host renewal trap is real and it is the entire business model there: the introductory year is a customer acquisition cost and the renewal is the product. If your client buys direct, they will get the introductory price, forget, and then ring you angry about a number you did not choose.

Check your withholding before you conclude the money vanished into coffee. My "$1,900 a month" second job turned out to be about $1,500 a month once I owed the shortfall in April. I was mentally spending money that was always going to be clawed back.

Worth a preparer's time for one session if you've never had two W-2s at once.

Cut the number of inboxes too. Four capture channels means four things to remember to empty, and in practice you'll empty two of them. One digital inbox and one paper notebook is about the limit for most people.

Also check whether your rate changed rather than your usage. A lot of bills that go up have nothing to do with behaviour, and it's worth comparing kilowatt hours used against the same month last year before optimising anything. If the units are the same and the money is different, unplugging chargers will not help at all.

Price the insurance before you do any other maths. Get an actual quote for equivalent coverage for two people on the open market where you live, monthly, including the deductible you'd realistically hit. In a lot of cases that number is between $900 and $1,800 a month, and it turns the salaried role from $138k into something closer to $155k in real terms.

Then compare against the contract honestly: 25 hours at $95 is roughly $123k annualised if you never take a week off and it never gaps. Contracts gap. A five month SOW with a vague renewal is not five months of certainty, it's five months of income and then a job search.

On those numbers I'd keep the salaried job and be very deliberate about fixing the boredom inside it.

That's $16k a year of compensation you were about to hand back without noticing. Boredom is a real cost too, but it's a cheaper one to fix than health cover.