Data accuracy is vitally important for unions that want to improve the member experience. This has been doubly so over the last year, where many workplace reps have been unable to regularly check in with members face to face, and most communication has moved online.
But when a member changes their online contact details and forgets to update them with the union, it sends the union’s campaigns into a black hole. The member gets cut off from opportunities to get involved, and worse still, the union might not even realise it.
For this project, the Digital Lab worked with colleagues at RCM to test ways to check and improve contact data accuracy. With a strong focus on offering specialised professional training to members, accuracy of the union’s personal data on members is very important in helping them get the most out of their membership.
We agreed that this challenge could be tackled with a digital intervention, using design thinking methods. The results could help RCM’s membership team to plan their ongoing efforts towards this goal.
Discovery phase
We kicked the project off prior to the pandemic with a workshop to look at questions around the accuracy of member data.
Working with two consultants, John Chadfield and Eth Morgan, the initial workshop brought together six colleagues from different member-facing roles in organising, member support and membership data functions.
Using post-its on the wall, we discussed and mapped a typical life cycle of a member with the union, from joining as a student to retiring completely from midwifery. Alongside this, we mapped different interactions the union had with them, helping us see where data was being exchanged. Establishing a shared narrative for the member’s journey helped us to identify several pain points and opportunities for interventions to work on member data accuracy.
The workshop found that transitions between different stages of their members’ working lives could represent a high risk of losing data accuracy. There was a particular concern in the transition from student membership to professional qualification and first job. Members’ details may become invalid when they leave their university course, and the university-provided email address that comes with it.
We had to park the project in 2020 as the union responded to the pandemic, which demanded urgent support for their members working across the NHS. More recently though, we have been able to return to prototyping one of the two ideas that came from the workshop.
Prototype phase objectives
- Test out our assumption that there is a digital solution to improving member data validity
- To design and develop the lightest, least-risk, intervention to use for this test
- Deploy and monitor/analyse the intervention
- Report back to RCM on results, learnings and options to replicate benefits at scale.
Hypotheses we wanted to test
- Members’ mobile numbers are the strongest (most reliable) data point.
- SMS could be a way to tackle the gap for members moving from study to work.
- A prototype intervention could improve the quality of member data by texting members to confirm or self-update contact details.
- Making this process as simple as possible from the user’s perspective will increase the proportion of members completing.
The purpose of the prototype phase is to produce a usable digital intervention that can be used to test our hypotheses, and be ‘handed over’ to the union, so that if the project proves successful, they have the option of expanding the number of members with whom they deploy.
As always with digital product design, the ambition is to keep the fidelity of the prototype as minimal as possible in order to convey the value of the proposition, rather than sink time and cost into a bespoke solution that may not be fit for purpose.
This kind of test lends itself well to prototyping. There are many third party services available, which can cheaply replicate the user-facing functionality of more complex integrated systems. Working at smaller scale also means additional staff effort can be flexible in covering any gaps in the system, so fewer eventualities need to be built in. If the prototype succeeds and the union decides to take it up, they may choose to build it more thoroughly, or integrate it better with their systems.
Prototyping design
A prototype workflow was created using ‘off-the-shelf’ technology services that allow a member’s contact details dataset to be processed into a two-way, automated SMS communication campaign.
The aim was to engage members within SMS to start with, bringing only those members who identified an issue with their email address across to a form to correct it. Moving users from one channel (SMS) to another (web) is fraught with risk, and so it was crucial to make that journey as easy and seamless as possible for RCM members taking part in the test.
Even though the technology deployed was only a temporary and partial solution, we were using live member data. This meant we still needed to give full consideration to the security and integrity of that data. A Data Protection Impact Assessment was carried out first, to identify potential risks, and data processing agreements drawn up, so all parties involved had a common understanding of expectations on them and how we could mitigate any risk.
The results were very positive and could point to other useful steps for the union to take, as well as for other unions to consider.
Dataset: 800 RCM members.
- Split into three test groups: 300 newly qualified members, 300 regular members, and 200 regular members who were direct to the union’s regular membership portal as a control group.
- The mobile phone number data was very accurate, suggesting our hypothesis could be right that mobile is a stronger contact point. When uploaded to CallHub, the software checks for valid and active numbers, and it only found two that could not be used.
Tech stack
- Typeform (web survey) – using “Plus” account, £49 per month.
- CallHub (SMS send and response) – Total CallHub spend on SMS for the project was just over £100. This included monthly fees for renting three numbers for members to respond to, and pay-as-you-go fees per SMS sent, either broadcast, manual or automatic response.
Design Considerations
- Users needed to be shown enough data about themselves to signal that this was a legitimate campaign/point of contact from RCM. Everyone is well used to receiving phishing messages and this can make them wary about interacting with requests for information.
- We used the member’s first name in the initial salutation, using CallHub’s merge facility.
- We showed members the email we had on file and asked if it was correct, rather than asking them to supply information up front.
- RCM website team created aliases on the union’s main URL (eg rcm.org.uk/update). These redirected to Typeform or Bit.ly, but gave users more confidence they would be interacting with the RCM than a generic service URL might have.
- The Typeform used RCM branding and referred to the RCM privacy policy to give additional reassurance.
- Design of the script and subsequent webforms needed to be simple and uncomplicated in order to maximise completion
- We asked only for a “yes” or “no” response to the first question, to remove any confusion on how to reply. A different response text was sent automatically by CallHub, depending on the reply.
- We set SMS response email alerts to let us respond manually to anomalous replies. We needed to factor in time to be able to respond promptly to any such messages, as people tend to expect faster SMS responses than via email. If we were running this again, the automated messages could be fine tuned to ignore some other more common replies and allow personal answers instead.
Conducting the process
- Rudimentary A/B testing was carried out to compare completion rates between users presented with the Typeform ‘Update’ solution, and the existing RCM Portal.
- For the Typeform, we ran separate tests for a group of 300 newly qualified midwife members and one of 300 longer term members. Around 200 longer term members were directed instead to the union’s regular member portal, which allows them to log in and change their details with the union.
Results
- 50% engagement across all groups
- This is the number that responded either YES or NO to the original SMS.
- Benchmarked against industry/charity campaigns, this is an exceptional figure (20% range being deemed ‘excellent’). It shows the union has strong recognition and engagement with its members.
- This confirms our hypothesis that SMS is the strongest contact datapoint RCM has on their members.
- The result was additionally impressive given that there had coincidentally been a phishing scam circulating with midwives, purporting to come from the National Midwifery Council, which the union had recently notified members about. This may have given some members pause, and indeed we had two responses querying this that we replied to manually to convince users.
- Of the ‘NO’ respondents, 45% moved from SMS to start the Typeform.
- This confirms hypothesis that if a suitably simple digital intervention were deployed, RCM members would use it to update their data.
- This is also a creditable response as moving people from one platform (SMS) to another (web) usually indicates a large drop off, as people tend to prefer to stay within that channel.
- If we were trying again, a return path purely in SMS could be tested to allow people who said no to simply reply with their current email address, without the need for the Typeform step. This may increase the conversions further.
- Completion rate of users finishing the typeform and updating their details was 85%+
- Average time to complete was just 17 seconds. This is a factor of having a minimal form. If we had introduced more data to update it would have depressed the conversion.
Group 1 – New-qualified members
- Of the 12 who used the Typeform to update details, 9 changed from university email to personal email, and three from a personal email to a new personal email.
- This validates our workshop’s insight that newly qualified midwives were forgetting to update their email address when leaving university, but keeping their mobile number.
- It also confirms our hypothesis that when members were presented with their old university email address, when pointed to the Typeform prototype they were motivated to update it for the union.
Group 2 – Full-rate members
- 12 members changed details. 4 from work email to personal. 4 from old personal email to new personal. 1 from work email to a new work email. 1 from personal email to new work email.
- Interestingly two of the old personal email addresses changed were from what looked like a family or partner’s email address to one that looked like the member’s own. These may be members who have not updated their email in many years, who may not have been such active internet users when they initially registered it. This means they may not have been seeing union emails that were going to an inbox primarily run by someone else. From this, it may be worth doing this kind of process with people who have not updated in a long time, as whilst the email may not be bouncing back, it may no longer be the best one.
Group 3 – Full-rate members sent to the portal
- Of this group of 200 members, 19 responded “no”, that their email address was incorrect. They were sent to the member portal.
- Of these 19, none updated their email address using the member portal. However, 3 of them did get in touch separately with RCM Connect during the period of the trial, and asked for their details to be updated, so we can count these as likely influenced by the SMS message.
- This would confirm the hypothesis that when working with less engaged members, it would be worth going the extra mile to make the process as simple as possible for them.
Lessons learned
SMS is a valuable engagement channel for unions
SMS as a channel is high engagement, trusted and good value for money. As personal contacts have moved onto messaging or social media, people are now more used to receiving SMS with service alerts from organisations and companies that they interact with (such as getting medical appointment reminders, or updates from the bank). Provided it isn’t overused with less relevant asks, this can work for unions too.
Where unions have the chance to gather SMS information, it will be a useful resource in engaging with members. For the RCM, gathering SMS details for more members in student and newly-qualified levels of membership will be especially important in providing a backup to email addresses.
There will however be additional considerations in asking for SMS information and processing it according to data protection regulations. Unions should consider this, and also only ask for the information where appropriate. Adding it in where it does not seem justified to the member will impact on completion rates for other engagements.
Personalising the ask is effective
When we showed members that the email we had on file for them was their old university address, the pilot evidenced that they were motivated to update it.
We believe the response would have been much lower had we merely used a generic “please update your information” message, without showing them.
Considering the member’s motives is also crucial. We overtly explained that the reason we were asking for accurate contact information was so the union could better support them when needed.
Simplicity is key
Well-designed, mobile-first forms play a big part in encouraging members to engage with interventions such as this. Once they reached the form, more than 80% completed it and updated their data.
Typeform is designed to work with mobile users, and so is a natural choice if you were to consider further campaigns of a similar nature. We wouldn’t recommend this intervention as a way to get people to sign up for an existing service, unless it had been tested to have a high completion rate on mobile.
With every new choice or page of information added to a journey, a number of users will drop out. There is always a risk that in adding more complexity, you end up causing a “lost cart”, where the user starts to complete the action but then drops out. This experience may hurt future click rates with that user if they expect other interactions you ask of them will be complicated.
Do the work for the user
Even with the simple methods used, creating this workflow and reimporting the data back into RCM’s CRM did create additional work. It would have been much easier on the organisation to simply direct people to the membership portal.
This is not to say that a facility to update information on the website is not important (It is, and all unions should work towards integrating their CRM and CMS in this way – effort in optimising this to be as easy as possible will also be well spent). However, that works in the context of an active user who is already motivated towards the task of informing the union. With this, we had to interrupt members and convince them to take time out from whatever they were doing, so we needed to put additional effort in to make the process easier for them – even at the expense of making it harder for the union (see our blog on BJ Fogg’s behaviour model for more on this).
The results suggest the extra work is definitely worth it. The roughly 4% of members messaged who corrected their emails will mean an ongoing 4% increase in deliverability (and hence open rates and clicks) for all the union’s communications and mobilisation by email.
If you’d like to discuss how you could run a similar project in your own union, or have insights into how this could be done more effectively, please do get in touch.