Pete

Disagreeing slightly with everyone telling you to work on the cohort. Four people is not a bad number, it is a small number, and two years is short. I would put the effort into one non university thing and one existing friendship maintained properly at distance, rather than trying to reconstruct an undergraduate social life that was an artefact of being eighteen in a building with two thousand other eighteen year olds.

The mistake I made was giving the helper my login instead of their own seat. No audit trail, replies that came from me, and when we parted ways I had to rotate everything and could not tell which of the last two hundred replies were mine. Give them their own account with their own name on the replies from day one, even if it costs an extra seat fee.

Don't argue about what was agreed, just show it. Put the original brief in one column and the current list in another, no commentary, and send it with a note saying the work has grown and here are the options. Then give three: deliver the original five pages at the quoted price and schedule the rest as a second phase, deliver everything at a revised price of X, or cut specific items to stay in budget. Almost every reasonable client picks one and the conversation stops being about blame within one email. The version of this that fails is the one where you explain how you feel about it.

They almost always do. Clients aren't trying to rob you, they just genuinely lose track of what they asked for.

Check whether the page lost internal links. This happens more than people think. Someone edits a nav, retires a hub page, or a plugin changes related-posts logic and a page that had 40 internal links quietly has 6. Crawl the site now and compare to any old crawl you have. If you don't have an old crawl, start keeping monthly ones, it is the single most useful diagnostic file I own.

There it is. Rebuild those links from live pages, not from an archive page nobody visits, and give it three weeks.

The going from competent to incompetent thing is exactly what I have been circling and not naming. Good to know it lands around month four.

so the signup number was never a signal. what is the equivalent number that would have been real?

Stop talking to your contact about it and start talking to accounts payable directly. Ask your contact for the AP email and the internal reference number for your invoice, framed as a logistics question rather than a complaint: 'can you send me the AP contact so I stop bothering you about this'. Nine times out of ten you'll find the invoice was never entered, or it was entered against a PO that doesn't exist, and your contact genuinely doesn't know because they've just been forwarding it. Once you're talking to the person who runs the payment run, you get a date instead of a reassurance.

Classic. Ask them now what the PO process is for the next job so it never happens again, while they feel slightly guilty.

Don't submit them for indexing one at a time, it doesn't do what people hope and you'll burn a week clicking.

Adding the cheap safety net that would have caught your two endpoints: a test that enumerates the router.

Mine walks every registered route, calls it with a member-level session against another org's resource, and asserts 401 or 403 or an empty result. Any new route is opt-in to the test by existing. It found three unguarded endpoints the day I wrote it, and it has found one every couple of months since, usually within an hour of the mistake.

It took an afternoon. It is the highest-return test in the codebase by a wide margin.

The part nobody says out loud is that your timeline is set by your advisor more than by you. If they want another chapter, you write another chapter, and no financial planning survives that. So alongside the money question, ask senior students specifically how long your advisor's students take, not how long the department takes. That number varies more between advisors in the same department than it does between departments.

Make sure the wording in the contract says capacity, not outcome, or you've just written a fixed price contract with extra steps.

This has bitten me twice. Now I always check the CRM for URLs in email templates before killing anything.

The line is what versus how. Tell them exactly what's wrong and why it matters, in as much detail as you like, and stop before the implementation plan. 'Your product pages are being outranked because there's no internal linking from the category level, that's the main constraint' is a diagnosis that makes you look sharp. 'Here's the linking structure and the anchor text pattern' is the work. Most people can't act on the first without buying the second, and the ones who can were never going to hire you.

The 2024 one pulls a decent tail around specific mistakes. So differentiate, probably.

That's the actual bug then. Most people price against an employed-hours model and then wonder why the maths never works.

One practical thing while you wait: go find where your audience already hangs out and be useful there. Not to drop links, just because the first hundred readers of a new site never come from search anyway.