Budgeting
Budgeting apps sell you a bank feed, and the feed is the hard part
The clever categorisation is not what you are paying for. You are paying for a reliable connection to your accounts.
This is less a set of instructions about budgeting tools than an argument, and it is worth saying so at the start.
The argument in brief
- Account connections expire and have to be reauthorised periodically.
- An app cannot see cash, unlinked accounts, or a partner's money.
- The tool matters less than whether you actually open it.
What the connection actually does
Most budgeting applications work by reading transaction data from your accounts through a permissioned interface that the bank provides. You are granting read access to a third party, which is a different arrangement from handing over your login credentials, and the distinction matters. Where regulated interfaces exist, that access is time-limited by design and has to be renewed by you at intervals.
Renewal is the most common reason an app quietly stops updating, and a stale feed is worse than none because it still looks current. Anyone relying on the tool should check the last-updated date on each account rather than trusting the headline balance.
Why feeds break
Connections drop when a bank changes its authentication process, when a card is reissued, or when consent simply expires on schedule. Some account types are supported inconsistently, particularly savings products, business accounts, credit agreements and anything held with a smaller provider. A broken feed on one credit card makes the whole picture wrong in the direction that matters, by hiding spending rather than income.
Applications differ enormously in how loudly they announce that a feed has failed, and some are close to silent about it. The practical safeguard is a monthly reconciliation against the actual statement, at least for the accounts carrying most of the spending.
What the software cannot see
Cash withdrawn and then spent appears as a single transaction at a machine, so that spending simply cannot be described by the tool. Money moved between your own accounts often registers as both income and expenditure unless transfers are correctly identified and paired. A partner's spending is invisible unless their accounts are linked, so a household picture needs either shared access or manual assembly.
Anything paid through an intermediary or a wallet shows the intermediary as the merchant, which obscures what was actually bought. These blind spots are structural rather than defects, and every conclusion drawn from the data inherits all of them.
Free, paid and what is being sold
Free tools are funded somehow, most commonly by referring users to financial products, and that shapes what the interface chooses to highlight. A product suggestion inside a budgeting app is advertising unless it is explicitly presented as regulated advice, and the two can look similar.
Paid tools remove that incentive but not the question of what happens to the data, which is worth reading the policy for. The strongest signal is whether the tool earns from your subscription or from your onward application to something else.
Neither model is disqualifying, but knowing which one you are inside explains a great deal about the notifications you receive.
Spreadsheets and the manual alternative
A spreadsheet has no feed to break, no data leaving your control and no product suggestions, at the cost of typing. Manual entry has a genuine behavioural side effect, because recording a purchase by hand inserts a small delay between the impulse and the record.
Its weakness is completeness, since manual systems drift as soon as a busy fortnight passes with nothing entered. A hybrid is common: automatic data for the accurate history, and a manual sheet for the decisions and the forward plan. The forward plan is the part no feed can supply, because it consists of intentions rather than of transactions.
Choosing on the only criterion that matters
Every mainstream tool categorises adequately, charts competently and exports its data, so feature comparisons rarely decide anything real. What differs is whether you open it, which depends on the interface, the notifications and how much correction it demands each week. A tool requiring ten minutes of tidying every week will be abandoned by the second month regardless of its capabilities.
Before committing, check that the data can be exported in a standard format, because that determines whether you can ever leave. This is a description of how such tools work rather than a recommendation of any particular one.
The takeaway
Judge a budgeting tool on whether its feeds stay connected and whether you open it, not on how many charts it can draw.
Write the number down before you decide. It usually decides for you.
Questions readers ask
Is it safe to connect an app to my bank?
Where a regulated open banking interface is used, the app receives read access rather than your credentials, and you can revoke it. Check that the provider is authorised in your country before connecting.
Why has my app stopped updating?
Most often the consent has expired or the bank changed its authentication. Reauthorise the connection and check whether transactions were missed while it was stale.
