I approached the DMA with a mixed bag of ideas, working through these with my Mentor we finally settled on running a series of experiments using post-show emails. artsdepot is yet to dip its toe into the pool that is post-show emails, which means we don’t have a baseline to measure success against. Initially I thought sales alone would be the measure of success, but over the course of the coming weeks, as I mulled over the experiment, I had two realisations:
- I needed to take a step back and ask… why do we want to send post show emails? What’s the gain? The answer: to encourage repeat attendance. So before starting we need to establish what artsdepot’s current repeat attendance patterns are. Then we can see if repeat attendance increases over the course of the experiment.
- When Sara Devine and Shelley Bernstein talked about agile working in their online workshop they gave an example of how they had tried to introduce recommendation cards for visitors at the Brooklyn Museum. If stewards saw visitors looking at artefact ‘x’ they would hand them a card suggesting they might also enjoy artefact ‘y’. Visitors did not respond well to this; it felt impersonal and left them confused. After hearing this I started thinking, what if post show emails have the same effect? How would we know? We assume they work but it’s not unreasonable to imagine a scenario whereby some people appreciate the email but many are left annoyed/confused. This made me realise that we also need to measure unsubscription rates as part of the evaluation.
There are a lot of variables that can apply in this experiment and a lot of questions that could be answered, e.g. could audience members from hire events be enticed to book for programmed shows? Would customers prefer a single personalised recommendation or an overview of numerous forthcoming events? Does there need to be a reason for getting in touch beyond marketing (e.g. a customer service survey)? How will these emails fit into artsdepot’s existing email schedules? How soon after an event should a post-show email be sent? Do recommendations work better for some audiences than others?
Which of these questions are addressed remains to be seen but for now the next job is to plan time and resources; to work with Spektrix to establish the baseline statistics, and to work out how we will run the experiment alongside our current email marketing.