Information sitting in a portal that somebody has to remember to download
The completeness of the record depends on somebody's memory.
Free, complete information sitting behind a login is only as good as the last time somebody remembered to pull it.
Not every business has this problem, and some should leave it alone.
Most of these can be done better. Whether it's worth doing is a separate question. What follows is one common form of the problem, described in general terms, because your version will differ. These are possible challenges, not a description of your business, and some should be left exactly as they are. If it sounds like your week, that's worth a conversation.
Sound familiar?
- Nobody has a schedule for pulling the export; it happens when somebody needs it.
- The same reformatting routine gets done by hand every single time.
- A pull once missed a weekend and it went unnoticed.
- Some of what could have been downloaded in March is gone by September.
What it looks like
Something important lives on a website the business logs into: a customer portal, a supplier account, a payment processor, a public records system. The information is complete and usually free. Getting it means logging in, exporting a file, and reformatting it by hand into something useful. Nobody schedules this. It happens when somebody needs it or remembers to.
What it costs
With most tasks in this guide, the item comes to you — an invoice arrives and waits, late but present. Here nothing arrives. If nobody remembers, there is no pile and no evidence anything is missing. The record ends up full of holes shaped like busy weeks, and any question that needs history gets answered from a dataset with unknown gaps, delivered with a straight face because the gaps are invisible. Worse, some of it expires: portals purge, and information free in March can be gone by September. That is the only kind of backlog that cannot be cleared by working late.
Where it goes wrong
- The export is reformatted by hand every time, identically, from memory.
- Date ranges overlap or leave gaps when pulled manually, and neither is visible afterward.
- One person holds the login, so the task belongs to whoever has access rather than whoever needs the data.
- The file becomes the record. Exports pile up in a downloads folder that quietly becomes a system nobody else can navigate.
What a better version looks like
Collection happens on a schedule instead of a memory, and keeps going whether or not anyone is thinking about it. Each pull knows what the last one covered, so ranges neither overlap nor gap. The reformatting becomes part of the collection rather than a routine somebody repeats, and the original file is kept as evidence of where it came from. The most valuable property is boring: the record becomes continuous.
Questions worth asking about your own operation
- How much of your history exists only because somebody remembered to pull it?
- When was the last time a date range was accidentally skipped?
- What in this portal expires, and on what schedule?
- Who holds the only login?
If this sounds familiar
Bring me the version you actually have. I'll learn how the process really works before I suggest anything. If it isn't worth changing, or isn't a fit for me, I'll say so. Talk through a problem