A lot of the time there are good reasons, but the funding structures of this kind of large-scale science also generate some perverse incentives. No-one's employed full-time to "be a scientist" at the kind of scale that a space mission takes - at best some of the people will be tenured researchers, but their loyalty tends to (reasonably) belong to their specific institution. Usually at least some of the people involved will be commercial contractors who will then move elsewhere once the money goes away - and it will, as soon as the probe is launched. Finally, science tends to be funded on a grant basis, and these grants are always for specific things.
So imagine someone working on the Foo Mars Orbiter. They're probably employed by the University of Bartown, and their work is being funded by an NSF grant for the Foo Mars Orbiter project. Neither of those bodies is interested in making something reusable for another satellite ten years down the line: the university may well not be involved in the next one, and they're unlikely to get much academic prestige from having their designs reused, they'd rather their professor spend time making their paper nicer and getting it in better journals. Meanwhile the NSF wants its grant to be used for the grant's specific purpose, not as a general slush fund, and will be hostile to any expenditure that's not directly related to getting the Foo Orbiter into orbit. Even if the team did put together something reusable, the team's going to be dissolved as soon as it launches (with a skeleton crew remaining to check when it arrives) or at very best once the primary mission completes, so no-one's going to be in a position to handover the assets to the team making the next satellite. If you've ever tried to reuse some code that was written two years ago by someone who thought it was a one-off and has since left the company, you get the idea.
So imagine someone working on the Foo Mars Orbiter. They're probably employed by the University of Bartown, and their work is being funded by an NSF grant for the Foo Mars Orbiter project. Neither of those bodies is interested in making something reusable for another satellite ten years down the line: the university may well not be involved in the next one, and they're unlikely to get much academic prestige from having their designs reused, they'd rather their professor spend time making their paper nicer and getting it in better journals. Meanwhile the NSF wants its grant to be used for the grant's specific purpose, not as a general slush fund, and will be hostile to any expenditure that's not directly related to getting the Foo Orbiter into orbit. Even if the team did put together something reusable, the team's going to be dissolved as soon as it launches (with a skeleton crew remaining to check when it arrives) or at very best once the primary mission completes, so no-one's going to be in a position to handover the assets to the team making the next satellite. If you've ever tried to reuse some code that was written two years ago by someone who thought it was a one-off and has since left the company, you get the idea.