From January 2021 until July 2021, a group of members of the POP consortium, DiRAC, the N8 CIR, Durham University’s Institute for Computational Cosmology and members of Computer Science organised a series of seven workshops around performance analysis and optimisation. The series covered a variety of tools, featured different invited speakers, and addressed inter-node, intra- node and core-level runtime efficiency. In contrast to established workshops that usually run en bloc over multiple days with individual participants, we organised a series of one-day workshops and invited teams of participants. Each team consisted of multiple (3–7) scientific software developers working together on one code base. This report summarises the workshops as well as feedback gathered at the end and throughout the series. It discusses interesting lessons learned w.r.t. the status quo of the codes/teams that signed up for the series, it summarises impressions regarding the training status and needs of the participants, and it relates these take-aways to current discussions around the training of high-performance computing application developers in the UK.