Since everything uses memory space to run, the way the memory is organised is highly important. Memory structure (or memory architecture as its also known) is the organisation of how the electronic data storage will be accessed and stored in the fastest, more reliable way possible.
Data in memory is manipulated by the CPU when it performs the calculations and movements necessary to run applications. In order to make this work as fast as possible, most computers employ a scheme of memory hierarchy that puts the most-accessed, or most-recently-accessed data closest to the CPU. However, this often requires a compromise with memory size though, as you cannot store the entire physical memory unit close to the CPU, so there will be cases when the data the process is searching for isn't instantly to hand and must be searched for more slowly. Memory is stored in a very incremental way. The most common unit of storage it a byte, which is 8 bits. Each of those bits is a 1 or a 0, and those bytes altogether make up a binary string that the computer can read quickly and easily for its applications.
Memory also allows the computer to store results of CPU operations for larger, more meaningful chunks of results. If we didn't have a as much memory as most computer builds today do, results would have to be put out instantly and would make little sense to the human reader. Also, the kinds of operations possible would be very much more limited and inflexible. Calculators are examples of computers with small memories that carry these limits, and it's clear how much more advanced computers are, and need to be. The extra memory is truly justified.