Synaptic changes driven by Hebbian plasticity are unstable and require compensatory mechanisms to homeostatically control neuronal activity. Synaptic scaling regulates the synaptic efficacies to bring individual neurons to a set firing regime [1]. However, the timescales of synaptic scaling are too slow to counterbalance the rapid growth due to Hebbian mechanisms [2]. Heterosynaptic plasticity that operates on timescales similar to Hebbian processes can prevent runaway potentiation [3] but its ability to homeostatically control activity is uncertain. Here, we introduce aggregate scaling, a simple theoretical model for synaptic scaling based on transit dynamics of synaptic building blocks, such as AMPA receptors. Competition over these building blocks among synapses ensures that activity-dependent changes are counterbalanced by heterosynaptic modifications [4]. The total amount of available building blocks is homeostatically controlled to stabilize neuronal firing. Using our model and experimental data [5, 6], we establish a relationship between the timescales of synaptic scaling and the activity sensor of the neuron that allows for stable convergence towards the firing rate set point while avoiding damped oscillations. We simulate a leaky integrate-and-fire neuron and a recurrent spiking neural network with conductance-based synapses. Excitatory synaptic weight modifications are realized as activity-dependent plasticity [7] and activity-independent processes [8]. A simple normalization scheme implements heterosynaptic plasticity. The firing rate of individual neurons is regulated with a control law that responds to persistent activity changes. We show that in both single neuron and network simulations aggregate scaling ensures stability, regulating firing activity at long timescales while allowing for transient activity changes, such as those induced by stimulation. Overall, aggregate scaling emphasizes the necessity for multiple timescales in homeostatic neuronal regulation. It overcomes non-realistic assumptions and limitations of existing models, such as access to non-local information, and provides a simple theoretical framework for studying homeostatic synaptic scaling.