The first is broad matching without negatives. Platforms interpret your keywords generously, which means you pay for searches that share a word with what you sell and nothing else. Open the search terms report, which shows what people actually typed rather than what you targeted, and add anything irrelevant as a negative keyword. Do this weekly at first. Most accounts find a meaningful share of spend going to terms the owner would never have chosen.
The second is sending traffic to the wrong page. An advertisement offering something specific that lands on your home page asks the visitor to find what was advertised, and most do not. A page continuing exactly the thought they had when they clicked converts several times better, and running advertising without one is the most common reason a campaign appears not to work.
The third is running without conversion tracking, which is not waste in itself but guarantees you cannot identify the other two. Research on small business accounts consistently finds a large share have no conversion tracking configured, which means the spending was unmeasurable from the first day and every conclusion drawn from it was a guess. Verify it by submitting your own form rather than assuming installation means working.
Beyond those, check the settings that default against you. Whether your campaign is showing on partner networks you did not choose. Whether location targeting includes people merely interested in your area rather than in it. Whether ad scheduling runs at hours when nobody who could buy is awake. Each is a checkbox and each quietly consumes budget.
Then set a stopping rule before spending anything further. What result would justify continuing, what result means stopping, and the date you decide. Businesses waste more money continuing than starting, because stopping means admitting the earlier spending produced nothing.
Check the geography of who actually clicked, not only what you targeted, because the location report frequently reveals spending in places you have no interest in serving. This happens through the default setting that includes people showing interest in an area rather than being physically present, which is reasonable for tourism and wasteful for a local trade.
Pause rather than optimise when a campaign is clearly not working. The instinct is to adjust, and adjusting a campaign built on a wrong premise produces a better version of the wrong thing. Stopping, working out what the premise should have been, and rebuilding costs less than months of incremental improvement to something that was never going to work.