When you work at a company that has a problem like this, the natural reaction is to go fire somebody in the IT department, and then identify "what went wrong".
Another option is to analyze how customers truly behave when you take the website away.
- Measure the anger on social media. Does social media anger correlate with sales declines?
- Measure online demand, by hour. You lose sales when the site goes down. Do the sales come back? If the sales rebound and exceed plan in subsequent hours, making up for what is lost when the site is down, then you have to re-evaluate every single attribution assumption used in your business model.
- Measure retail sales, by hour. Did anything happen in-store when the website went down? If the answer is no, then you have to question if the website has any role in driving retail sales. If the answer is yes, then you have to evaluate just how much credit the website truly deserves.
- Did the merchandise mix change when the site went down? In other words, if you cannot advertise various items online, does what the customer purchases change?
Pundits beat companies up when something goes wrong.
Opportunists use these situations as a chance to learn.
Do the latter, folks.