I'll assume that your KPI shows a significant increase from the first showing to the second. I'll also assume that the ONLY thing between the two measurements of the KPI is the MaxDiff.
Even then we cannot conclude that the MaxDiff itself is causing the increase. It could be that respondents are answering the second KPI measurement differently because it's later in the survey and that having questions in between the two measures is the cause. If you had another cell in which you had about an equal number of questions as the MaxDiff has (but not the MaxDiff) intervening between the two KPI measurements, and in that cell you did NOT see the lift, then maybe you could attribute the lift specifically to the MaxDiff and start to rule out that it results not just from order or fatigue effects.