That is because pairing helps mitigate risks (human error), when you're being paid to build software no client like to pay for bug fixing, they would rather pay

That is because pairing helps mitigate risks (human error), when you're being paid to build software no client like to pay for bug fixing, they would rather pay a bit more to make sure there are less errors (or at least appear more bug free). If you're a bootstrapped start-up, it is not an efficient use of time to be pairing 100% of the time. We tend to only pair on 'core' parts of apps or 'hard' problems, or when someone needs another pair of eyes on something (two heads etc), otherwise clients budgets are quickly exhausted on pairing for pairing's sake.