It could just be a mathematical thing. For example, in my workplace pension, I can see the employer and employee contributions separately. Despite buying the exact same fund on the exact same date, the employee part has a % difference a little bit higher than the employee contributions.
I put in all the contributions into Excel to double check, calculating the units myself, which were correct and got the same result. When I adjusted the contribution levels in Excel as a test, the performance also changed (just by a few bps), even though the fund prices were the same.
It seemed to be a rounding thing when buying units. For example, if you buy £100 of a fund at £2.453 and they give you units at 40.7764 (rounded to four decimal places). If you buy £240 at £2.453 on the same day, you get 97.8394 units.
Later the price of the fund goes up to £2.906. Your first 40.7764 units are now worth £118.50. Your second 97.8394 units are now worth £284.32.
So £100 has become £118.50, or an increase of 18.50%
And £240 has become £284.32, or an increase of 18.47%
Obviously these are just example amounts I made up, and the difference is tiny in this example. But in real life with larger varying amounts, different rounding mechanisms, equalisation, how often you buy them etc differences could add up?
Although you did say *markedly* different, so maybe worth some more investigation.