A mitigation strategy here is to delay implementing any new feature for a single customer, in order to query your other customers as to whether they need the same or similar feature. Oftentimes you will find other customers who are interested, even if their exact needs are slightly different from the first customer's request.
Best case scenario you get to build the feature once, based on a spec that fits the needs of a larger pool of customers. This reduces the overhead of more customers inevitably requesting similar features weeks or months down the line.
Worst case scenario you still wind up releasing a feature that only one customer is going to use. At least you've tried your best to fold more customers' needs into a single unit of work. Better to have asked and found no takers than to have not asked and wound up with repeated requests for similar features once it's too late to combine the work.
Best case scenario you get to build the feature once, based on a spec that fits the needs of a larger pool of customers. This reduces the overhead of more customers inevitably requesting similar features weeks or months down the line.
Worst case scenario you still wind up releasing a feature that only one customer is going to use. At least you've tried your best to fold more customers' needs into a single unit of work. Better to have asked and found no takers than to have not asked and wound up with repeated requests for similar features once it's too late to combine the work.