BECTU (the entertainment industry sector of Prospect) worked with TUC Digital in late 2017 to pilot a new way of joining a union for the London Visual Effects (VFX) sector.
Read on to find our more about what we did, what happened, and some of the useful lessons we learned as part of the pilot.
The campaign sought to recruit a burst of new members (a target of 50%+ membership in the big four companies in the sector, in order to pursue recognition) that would be attempted in a one month time period.
By showing the growth in membership, the project hoped to achieve a “Kickstarter” effect, helping to create enthusiasm and a sense of momentum towards a goal.
The software was developed with significant input from existing members of the VFX branch along with staff from the TUC and BECTU. It was designed by Sam Jeffers of product design consultancy The Shop, and built by freelance developer Tekin Suleyman.
The campaign fell short of the 50% target, but succeeded in nearly doubling the size of the branch, recruiting hundreds of new members, and repaying the investment an estimated ten times or more.
- To achieve 50% membership within the industry, leading to a recognition agreement for people working in VFX.
- To raise awareness of the union and the common workplace issues faced by VFX staff.
- To test the concept and software, to publicise it across the union movement with the goal of identifying improvements and further opportunities for pilot cases in other industries.
VFX industry background
- An estimated 6,000 people work in various VFX roles in London – one of the 3 international hubs for the industry (alongside Hollywood and Toronto). Half of these workers are estimated to be working in four large employers, with others spread between many much smaller ‘boutique’ VFX houses.
- It’s relatively well paid work. Salaries range from mid/late £20k’s for a junior, to £60k for a mid/senior level staff member.
- Many, if not the majority of, staff are employed on a contract basis, usually for the duration of the movie project they’re working on. People frequently move between companies.
- The industry is project based, making timing important. We had to choose a moment when the big four companies were fully staffed.
- The hours are long. When deadlines start to bite, staff often find themselves working 80 hour weeks, with evening and weekend working the norm.
- A very high proportion of the workforce are from the EU and the rest of the world. Many are concerned about the impact of Brexit on their career and choice of home.
- Concern that organising might make the industry less competitive and drive business abroad is widespread.
- In comparison with other parts of the movie industry, there is very low union penetration within UK VFX. Canadian VFX is already much better unionised for example.
- However, there is an active and established branch with hundreds of members and a committee of highly active reps.
- Reps and members are unused to putting their heads above the parapet and visibly organising. There is an unspoken fear of reprisals and black-listing in such a fluid industry.
The campaign therefore had some strengths to build on (the existing branch, the concentration of the employers and the salience of some of the issues faced by workers) and some weaknesses to overcome (poor conditions being an accepted “norm” among VFX workers, contract work, fear of what unionisation might do to the industry, little precedent for workers organising themselves).
This project built on a “Kickstarter” recruitment concept developed during a TUC design sprint in December 2016. Given the lack of union access to workplaces and the reluctance of members to out themselves, the kickstarter model seemed a good fit.
Workers were presented with a target (50% membership in their company) and two joining options. They could join immediately – with a first year discount on BECTU’s normal membership rates. Or they could register their willingness to join only if the target was achieved.
Over the month campaign, the home page was updated to show membership growth. Joiners and deferred joiners received email updates to motivate them and encourage them to share the joining page with their friends. Private sharing links for Messenger and WhatsApp were provided, so members didn’t have to share on public forums.
This mechanism had several advantages:
- Prospective members could feel more secure that they weren’t going out on a limb, and would only be part of something if it looked like it would work.
- We were able to harness people’s enthusiasm and curiosity in how the campaign was going, to suggest that they became recruiters themselves, given they were more vested in the campaign’s success.
On its main objective, the campaign failed. We didn’t get to the 50% target in any of the big four companies.
However, the Kickstarter mechanic of the campaign looked to be successful. Over the month, 28% of joiners chose the ‘join now’ option, with 72% choosing to ‘join if we hit the target’.
Following the campaign, the branch quickly phoned round the deferred joiners to explain what had happened. The message was that whilst the campaign had not reached 50%, the hundreds who had pledged to join would make the union a much stronger force in the industry. Anyone who still wanted to join was offered a first year discount rate if they still went ahead. This tactic had a conversion rate of approximately 30% of deferred joiners choosing to join up anyway.
The overall implied rate of joining for those joining immediately, and pledging to join later is 50% (28% + (0.3 * 72%)). Obviously for a successful ‘kickstarter’ type campaign, this would be 100%, but even for an ‘unsuccessful’ campaign, the uplift from the ‘join later’ option seems to be significant.
Doubling the size of a union branch is no mean feat, but members will not benefit to the same extent as they would if union recognition were achieved. The new membership may prove to be the raw material for a push in that direction, if a relationship with them is properly cultivated.
Estimating an average 3 year membership period for those that joined during the campaign, the overall return for the union in running the project should be over 10 times greater than the investment by BECTU and the TUC.
From a technical perspective, the software worked well. It was always available and there were no complaints about its functionality. The questions asked of users were well understood (only a few instances of people just putting their first name, or trying to avoid including their mobile number).
95% of people who started the joining process completed the task.
Across the campaign, the average conversion rate from visiting the landing page to joining was 4.5% (with a high of 14.5% of visitors converting on the penultimate day of the campaign).
In future iterations, the following would be beneficial:
- Editable copy, images and layout on the landing page.
- The ability to create multiple landing pages – e.g. on a per company or per location basis.
- A ‘download CSV’ script that any admin can run to give them access to the data they have permission to view.
- The ability to customise/update meta-tags, such as the Facebook OpenGraph tags.
We received some questions about the mechanic (“What should I do if I don’t work at one of the big four companies?”), but overall we believe prospective members understood it.
The “count”, and presenting progress to members proved a big problem. Working across an entire industry, then within four companies, made presentation hard, and misunderstandings were common, both within the campaign and among prospective members. We made the decision not to present live numbers about each company.
A lot of effort was made to make the landing page appropriate to the audience, by involving activists directly in its design through a participative design workshop. The name VFXassemble was devised by reps, who also settled on matters like colours, messaging and the content of key text elements.
There could have been greater investment of time and money in the visual assets we used. The look was quite home-made and the photos from campaign stunts often didn’t do anything to show a growing movement of people working towards supporting each other at work. There were a few comments on social media about the lack of visual appeal. Given this group of workers were highly design conscious, this was a stumbling block for some.
We conducted limited A/B testing during the campaign using Google Optimize. However the changes we made – removing the video and changing the number of buttons and language used on them – made no significant difference in click through or bounce rates.
Changing the copy and design wasn’t easy, requiring developer support as it was hard coded rather than user editable. This had been an agreed compromise to reduce cost and complexity of the software product, but future iterations should probably include greater direct campaigner control over content.
Awareness of the campaign was the most significant factor driving people to join. Direct traffic and organic search led to just under 65% of the total sign ups. This suggests that promotion of the campaign name and URL (possibly through the leaflet handouts, possibly through word of mouth) was the most effective single source of signups.
The days when high profile leaflet handouts were organised as workers arrived the mornings (launch day, plus the Tuesday and Thursday before the deadline) saw the highest levels of traffic to the website.
For these handout days, people in green body suits (to echo the green screen kit often used in VFX) stood with organisers outside the four main facilities. These stunts were clearly useful in raising awareness of the campaign. They may also have been useful in creating social proof that joining was “ok”, with more people feeling able to talk to one another within the facilities. There was feedback from some that people clowning in green suits was a bit naff, but it also helped workers identify that the activity was clearly aimed at them, rather than the other workplaces around them, and most people interacted in a very positive way.
The green suit activity also helped the campaign create visual content, which was then used by the campaign, and by the key industry publication, FXGuide.
In the final week of the campaign we sent a lot of email to existing members and new signups. These focused on getting people to talk to one another about joining. These didn’t seem to drive many new signups, though will have contributed to the awareness of the campaign and the sense of urgency as the deadline approached.
Email was useful in kicking off awareness of the campaign, but produced relatively few visits to the website. As well as the existing branch members, we had two separate lists that we used to talk about the campaign – a pre-registration list that we got industry influencer supporters to circulate in a teaser campaign, and a segmented list of VFX workers from a petition we ran about an issue of concern to them (securing credits on movies – something they often miss out on) These lists weren’t particularly productive though, and the teaser campaign had a significant proportion of people joining internationally, who weren’t relevant for the recruitment drive.
Mailchimp was a perfectly adequate tool for managing the email campaigns. It was easy to create, segment and test messages. Emails to joiners during the campaign were popular, with over 50% open rates, and worth focusing on to build a sense of excitement as the campaign built.
Facebook drove the largest proportion of traffic to the website (43%), though the conversion rate was slightly lower than for other channels. The Facebook page for the VFX union grew significantly during the campaign from 494 likes to 830.
Paid advertising on Facebook was used, but we didn’t gather enough information from its results to help focus it during the campaign. Whilst this was revenue positive, by targeting at people who claimed to be employed in the 4 main facilities, we were restricted to a small audience and unable to buy a significant number of Facebook ads during the time available.
A Reddit AMA (Ask Me Anything – an open Q&A session) was very successful, with 79 comments and 180 net upvotes. It was useful in two ways. First, it was an opportunity to put a human face to the campaign (key activist Joe Pavlo) and answer genuine questions from prospective members. Plus it also drove signups in its own right from a forum in which VFX workers were more active.
Twitter usage was muted, with 196 people taking part. Half of the Twitter activity came on launch day, with another small bump in the 5 days before close. A significant proportion of this was from union or industry influencer sources, rather than from members who were less inclined to be public in support.
Collaboration and ways of working
We worked on the principles of a content plan with the branch and BECTU/TUC staff, but it needed stronger co-ordination. Splitting the tasks allowed more to happen but meant people focused more on their own bit than the wider campaign. Individual work often slipped without a central product owner role holding it together.
We tried to use Slack to co-ordinate campaign activity between meetings and between the organisations, contractors and key activists. It was generally helpful, allowing discussion, sharing and approval of campaign plans and assets, but not everyone was used to using it and as different groups retained other communications methods it never achieved the critical mass it needed.
Overall this pilot seemed to be a promising tactic, but there are things a union should at least consider before attempting to run something similar.
How campaigns such as this earn and activate a greater number of face to face conversations will be key to their success. Ideas might include a greater number of leaflet handouts, more off-site meetups, creating videos, training and resources that show how to have conversations with colleagues and better anonymous tools to help people share the campaign.
To make email more effective in mobilising existing support, it would be worth investing in building up the email list ahead of the campaign launch, preparing existing members to play a role.
We ran no search advertising or remarketing during the campaign. Both would likely prove revenue positive if the resources were available.
We could have done more to prepare branch members to support the campaign directly. They contributed a lot of ideas and ways that we could use their talents, but it was harder to get ownership of the organising work of the campaign. A future campaign should involve a training focus, to help activists make the most of the campaign period and understand how they can talk about the union more effectively.
It would be good to ensure there is a central point of authority for the campaign, with clear capacity for the entire period of the campaign and time leading up. This could be a product owner drawn from union staff, or an external consultant given the responsibility and authority.
It would also be very useful to hire or appoint dedicated digital campaign support for the campaign, focused on a content calendar for email and social, and providing consistent support to back up the other parts of the team. If the time period is going to be this short then make sure to have at least the first half of emails and written content produced and signed off prior to launch.
Greater clarity around numbers would make the campaign more unified within itself. Maybe moving to a model around absolute numbers, rather than basing it on an uncertain starting proportion would be more effective. This technique might also work for building small teams across a chain employer, as well as building overall density.
We’re looking to develop some of the lessons from the pilot with other unions in different organising contexts, and hope to run further pilots during 2019, so watch this space. And do get in touch if you’ve got ideas.
And the BECTU VFX branch is doing well, with new activists taking a leading role in their upcoming campaigns alongside the existing committee. There will be plenty more to see from them in 2019 too!