Metal-oxide based Electrochemical Random-Access Memory (MO-ECRAM) has shown unique potential as a nonvolatile element for analog in-memory computation of deep learning tasks. Using a specially designed interdigitated device geometry, we investigate transient effects of MO-ECRAM and correlate them with programming speed, read speed and read-after write speed. Programming speed is shown to exponentially increase with programming voltage. Read speed reached the ns range, while read-after-write delay can be limited by decay of write transients in the studied devices. Two mechanisms of channel modulation were found; a prompt field effect and a field-induced memory effect. The charge control of the prompt effect was vastly greater than that of the memory effect. So to reduce and mitigate transient impact, we discuss both device improvements, and learning algorithm engineering strategies.