Are you an AMA member? please login

Personalised vs Basic #DigiLab

Danielle McLoughlin of Hull Truck Theatre talks us through her first experiment on Digital Lab and identifies four key challenges she has faced. 

I applied for a place on the CultureHive Digital Lab in the hope of gaining additional knowledge and support to develop the ideas and experiments that I wanted to implement at Hull Truck Theatre. I knew that the opportunity to work with a mentor and other fellows would be invaluable and allow for lots of interesting, cross-departmental discussions about best digital practices, which would then help contribute to the company’s longer term digital goals.

I had my first call with my mentor, Tom, and we discussed the various experiments I wanted to work on throughout the course. We talked about making sure that my experiment/s had clear aims and could be measured, and about making sure that each experiment contained limited variables so that we could analyse the outcomes effectively, getting meaningful results.

Tom really took me back to school with all of his talk about variables, action reflection cycles and methods of measurability, and following our conversation I had a much clearer idea of what my experimental goals were, and I was prepared to adapt to new ways of working.

Experiment Number 1

Goal: To encourage repeat attendance from customers, within an hour’s drive time, who booked tickets in 2017 (Hull City of Culture year) and have not returned since.

Experiment: To create an Oliver Twist (Christmas 2018) email split test and find out which type (Personalised vs Basic ‘What’s On’) encourages the most interaction and conversions.

Part One

  1. Extract all customers, within an hour’s drive time, who attended the theatre in 2017 but not since.
  2. Randomise these customers and separate in to four Groups (A, B, C, D) ensuring that there is an equal number of customers who booked for last years Christmas show, in each group.
  3. Email GROUP A a general Oliver Twist e-shot, including image, synopsis, dates, times and general booking information.
  4. Email GROUP B a personalised Oliver Twist e-shot, thanking them for being part of our ‘Year of Exceptional Drama’ and contributing to the City of Culture events, plus the above details.
  5. Analyse the interaction and conversions from both emails and make a conclusion about which email content was most successful.
  6. Use the ‘successful’ e-shot (let’s call this ‘e-shot X’) to then generate the next step of the experiment.

Part Two

  1. Email GROUP C with e-shot X.
  2. Email Group D with e-shot X, adding the show trailer, an animation or other visually engaging content.
  3. Analyse the interaction and conversions from both emails and make a conclusion about which email content was most successful.
  4. Use the ‘successful’ e-shot (let’s call this ‘e-shot Y’) to then inform how we might engage with lapsing bookers, digitally, in the future.

Following this my intention would be to carry out a similar experiment for out of town visitors that attended in 2017 but have not been back since. In terms of content this might be a simple show e-shot vs an e-shot with additional visitor information such as hotel links, packages, etc.

There have been numerous hurdles during the early stages of this experiment but whilst they have been challenging to overcome they have raised some valuable discussion points and issues that need to be addressed.

Challenge 1
Once I had planned my experiment I realised that to be able to analyse customer engagement effectively I would need to teach myself to use Google Analytics. I took part in the free online Google Analytics Academy and picked up the basics quickly. I also researched Google Analytics Tag Manager, as Tom said this would be useful and it will allow me to track which e-mail my web visitors had been referred from.

Challenge 2
I was eager to start setting up some filters and goals, on the Hull Truck Theatre live Google Analytics site, but the theatre didn’t have admin rights to its own Google Analytics account which limited the functions I could use.

Challenge 3 
The company who had admin rights for our account was an old OLD web host, who we no longer deal with and I needed to contact them to try and get admin rights. Luckily, we do still have a loose relationship with our old web host and he was extremely happy to work with us to resolve the issue. It took some discussion between myself, the old web host and our new web host (who we are just in the process of switching to – throwing another spanner in the works!) to find a solution but we now have admin rights to our Google Analytics account.

Challenge 4
When attempting to create a tracking tag so that I could see how many people converted and purchased tickets for Oliver Twist, I realised that the last seven pages of our checkout process have the same URI, This means that I can’t tell whether someone fully checked out or dropped off at the donation, delivery, billing or summary pages. I am now in conversation with our ticketing provider, who deals with the iframes in question, to see if we can make these URI’s more unique and thus accurately track the customer journey.


Image courtesy of Hull Truck. Theatre. Oliver Twist at Hull Truck Theatre, Flo Wilson, Samuel Edward-Cook and Lauryn Redding. Photo by Sam Taylor.

What we do #Digilab

Millie Carroll of The Tetley talks us through her progress on the Digital Lab

In the last few months I’ve made this project involve our whole team, and recently we’ve finalised some big decisions and we’re finally in the process of launching the new ‘artists area’ of our website.

However one of the biggest decisions we’ve made after some research is that this isn’t going to just be an area about the work we’ve done with professional and exhibiting artists, but an area that highlights the ‘what we do’ more broadly. We’ll have first person narratives from workshops leaders, children who take part in our after school club, volunteers, and anyone else who we can think of that may have been impacted by the work we do at The Tetley.

After a few weeks of research we have not found any other museums or galleries that have an area like this. We’ve found directory style pages on websites where they list the artists they work with, alongside images, videos and quotes from their time with these people, but none of them are reflective, which is the main aim of what I want to achieve. I want feedback on how their experience went such as, “My exhibition at The Tetley has led to me showing work internationally…”.  I believe the public and potential funders will be more interested to read about this rather than ‘I’m excited to have my exhibition at The Tetley’.

I’ve been in touch with three people we have worked with so far. One had a major exhibition with us in 2017. The other two have been part of our Springboard Programme who lead workshops for families and young people at The Tetley and are also freelance practitioners. I sent around six questions which were very open in the hopes that the people we get in touch with have the ability to speak their mind, I’ve advised them to view these questions as ‘prompts’. The feedback has been great and I’m excited to get it online!

We’ve had an issue with our new website — we need to move it to a new server. The plan was to get these posts online by the beginning of March but sadly it has been pushed back. After speaking to the technicians, hopefully by the 18 of March I’ll have these first three posts online and we can start measuring engagement! The idea of building a new area of the site has been more complicated than initially planned. Looking back I have been perhaps naive in thinking it would be easy to make changes to our site… Hopefully this will all still be done in good time!

Read Millie’s first Digital Lab blog — Using the past to impact the future #DigiLab

Image courtesy of Liverpool Biennial 2016 © Joel-Chester-Fildes 

How to win at YouTube #DigiLab

Danny Evans of the Royal Shakespeare Company keeps us up to date with her Youtube experiments on the Digital Lab


We have 500+ videos on our YouTube channel. Although we create a lot of really great content, some of which does really well, I wondered if we were getting the most out of our channel.

I set out to explore this and to do something we’d never done before – create a piece of content specifically for YouTube that targeted the audiences on the platform that we wanted to attract, and to work out ways to make our channel work harder for us.

Before I could create the content, I needed to work out what it would be – who I was targeting and what would work well with them. I talked it through with my mentor, Seb Chan. He asked me a big question.

“What does success look like?”

Seb asked me to look at our YouTube channel and find out which video was the most successful. This sounds easy – so I’ll just go and do that now…

YouTube doesn’t have a ‘which one is best’ section in its Analytics. We look at stats for individual videos, but don’t often make comparisons. I rummaged around and learned all sorts of interesting and useful things about our videos. But which one or which format was best?

  • Is the most popular video the one that gets most views?
  • Is the most popular video the one that people stay watching for the most time?
  • Or is it some kind of combination of these?

We’ve long known that our ‘most popular’ video on RSC YouTube is the balcony scene from Romeo and Juliet. It’s no surprise – arguably the most iconic scene by the world’s most famous playwright, loved by people and taught to children all over the world. It was posted way back in 2011, but still consistently comes out as our most watched video – depending on how you define ‘most watched’…

The 2011 Balcony Scene video is 12 minutes long, and although hundreds of people have watched it from all over the world, most of them don’t stay to the end. On the other hand, our trailers for upcoming shows, which are often about 30 seconds long, receive fewer visits, but are generally watched through to the end. How do you compare a video that is 12 minutes long but people drop off after three minutes, with one that is 30 seconds long and they stay to the end?

I decided that the most popular video was probably some kind of a mash-up of these two figures. And in doing this I gave myself a world of pain (maths). We (I got help) looked at these stats:

  • Length of video
  • Average % of video viewed
  • Number of views

We put it all on a spreadsheet, combined the figures, and it spits out a relative ‘popularity’ value (z) that by itself is totally meaningless and bewildering (The Balcony Scene gets 70.47). But when we compare all the videos together the value starts to mean something. Ordering the spreadsheet by this column, gives us a list of our videos in popularity order.

What does success look like?

Looking at the videos at the top and bottom of the table to see what works well and what doesn’t work so well, this is what I saw:

  • Our top 6 videos are all trailers
  • Video #7 is a skit on the To Be Or Not To Be speech, featuring a number of celebrities
  • A recent behind the scenes timelapse comes in high at #11
  • All the other high-ranking videos are trailers or scenes from shows (The Balcony Scene comes at #17 out of 500)
  • With the exception of The Balcony Scene, all the top 20 videos are under 2 mins
  • All our lowest performing videos are over 3 mins
  • The majority of low performing videos are talking heads

With 500+ videos on our channel, anything that comes in the top 100 is doing OK. That’s what success looks like. And if we want to make it work, we should probably make sure it’s under 2 minutes long and not a talking head.

So now I just have to decide what I’m going to do (and win at YouTube)…

Image Courtesy of Stellar Quines ©


Shrinking changes for the Sex Pistols #DigiLab

Ruth Selwyn-Crome from the UEA discusses her early experience on the Digital Lab including getting to the heart of her experiments 

I joined the Digital Lab to experiment but to a certain extent the UEA Gig History project I’ve been working on for the past 23 months has been an experiment from the start – when UEA took a bit of a PR risk and decided to mark the anniversary of the cancellation of the Sex Pistols back in 1976 with a celebration of 63 years of gigs hosted on its campus. The project proved a hit with alumni, staff and the gig-going public in the region and continues to attract followers with its dedicated Twitter feed @UEAgighistory. The project website played its part well – showcasing material and guest blogs on a monthly basis – but now needed to be taken forward to a more permanent yet reactive state. I needed ideas and advice and, in the process, hoped that the knowledge I picked up along the way would also be useful for my colleagues here in the public events, alumni and conference teams as we strive to attract new viewers to our livestreams and recordings and look for new ways to share.

So far my time in the lab has been taking part in as many of the online workshops I can and so far they’ve all given me new tools to play with and share, as well as some advice. In addition I’ve had a great talk with my mentor – Daniel Rowles from Target Internet and have been given some doable homework which has helped to manage what could be rather a daunting and distracting challenge. Takeaways from the workshops so far have been:

Shrink those Changes

Initial visions for a more reactive website included a range of possibilities, from offering site visitors the chance to compile their own yearbook of favourite gigs to entering a digital museum, with rooms and podcasts and film and a whole “immersive” trip down memory lane…

The first workshop I attended was “Scrappy Experiments”. A big takeaway message from that was to “Shrink those Changes” and make them new, different but most importantly manageable. I forced myself to go back to the heart of what the project has been all about – i.e. attending live gigs and having a great time and then being reminded of that great time. The list alone in a published book – had already proven incredibly popular.  So my first task was to get the list into a more usable format; 80+ pages of gigs, back to front, in a word document, peppered with tabs has been transposed into a new, more experiment-ready shape via Excel.

Don’t steal from yourself (steal from others)

Another lesson picked up from that first workshop was not to “Steal from yourself by comparing yourself unfavourably to others”. Task/procrastination no. 2 I’d set myself was to research what OTHER music venues were doing with their archives – out in the real world, away from Higher Education. I’ve found some great examples of what I really wouldn’t want to use and some surprising ideas from elderly institutions such as the Albert Hall who have a nifty online perpetual calendar/time machine – take a look here. There are some great archives out there too – you know, ones that actually list everything that’s ever been performed in the, e.g. streets of towns or in whole regions, while others concentrated on particular genres. I loved what the Roundhouse had done with their history, including interview/podcasts with staff and ex-staff from across the years. This form of celebration of the whole venue community was more what I was aiming for. Venues encompass the performers, the attendees and the employees and many volunteers after all.

Visualise, visualise, visualise those goals

I’m good at visualising – ideally I wanted a website that looked almost exactly like an online version of the Rip It Up – the history of Scottish Pop exhibition I’d visited in Edinburgh back in August. The exhibition installation ticked every box going with its black walls, costumes, surprises round the corner and loud music always slightly out of reach…and a juke box! Now I just had to find a template to play with…

Tooled up

Three more online workshops over the past months and I’ve been equipped with some great tools – especially to search for influencers on social media and find out what people are asking (‘Whatever happened to the Dead Kennedys?’). My colleagues in the digital marketing department have given me access to the gig history pages on Google Analytics. The Google Analytics workshop gave insights into the mind-blowing depth of analysis it’s possible to drill into. This will be useful to re-visit when I’ve set up my test website. I’m finding each workshop adds something to the last – e.g. User Journey with Target Media’s Daniel Rowles included a tip about a free Chrome plug-in for KLEAR which is proving useful already for identifying Twitter influencers in the sphere of Music. I’m also looking forward to using for advice about tests for my user group.

But first I need to finish my homework for my mentor Daniel Rowles. Daniel made me take a deep breath and asked me to identify three gigs which had a lot of memorabilia attached to them (done), then find a WordPress template to try out a few things on with some project contacts (kind of done). It’s been more difficult than I thought trying to find a ready-made template that looks like the inside of an old gig venue – with a sticky carpet and cigarette stubs – but I think I’ve identified a couple of candidates.

Watch this space…

Header image courtesy of Tower Bridge Guildhall School of Music Drama  © Paul Cochrane

See, Think, Overthink, Share. #DigiLab

Jade Joseph of Ideas Test discusses her initial concerns about being involved in the Digital Lab

Having permission to experiment in a work-based environment surprisingly feels very daunting, perhaps because it doesn’t come with any guarantees. Not to mention that for a relative newcomer to the world of digital marketing, there seems to be an entire language to learn and an avalanche of abbreviations to master. It felt inevitable that whatever ideas my fellow and I initially devised would shift and change throughout the workshops and mentor meetings. What I hadn’t prepared for was feeling a little bit lost when it came from moving an idea from a concept to something more concrete.

Before Digital Lab officially kicked off, it was easy to feel quite enthusiastic about my idea, which was to investigate the impact that some sort of regularly scheduled Instagram content might have on engagement. We have experimented with physical brochures in the past, we’re fortunate to have a website and a physical space in which to display details about our events but in true millennial style, I wanted to test whether dedicating time to creating content for Instagram would make a notable difference. There was still a lot to figure out but at this stage that felt okay. It was suggested that we try and narrow down our focus by the time we speak to our mentor to really identify the key areas that would be feasible to work on and to test with relevant results.

During the planning stage, I felt relatively optimistic. There’s the dual enjoyment of doing something creative and the challenge of getting an idea to actually work outside the way it exists in your brain. My first idea manifested as an Instazine; an aesthetic series of swipe-able images uploaded every fortnight to advertise our events and bring our audiences behind the scenes. I consulted and surveyed the rest of our team about the idea, loosely planned out what could be included in the first few issues, compiled thematic undercurrents to draw each instazine together and then, after our first mentor meeting, promptly lost all confidence in the idea.

“The subject of SEO had always felt quite alien and my knowledge of Google Analytics barely scratched the surface”

This was not entirely unexpected – as a world class worrier, a brief loss of confidence is nothing new! What was tricky was knowing what to do next. This is where the online Digital Lab workshops have been so beneficial. Being able to access the series of accessible online workshops has provided a lot of reassurance. They have been a great gateway into topics that previously seemed quite dense and tricky. The subject of SEO had always felt quite alien and my knowledge of Google Analytics barely scratched the surface, but the Digital Lab sessions were so informative that I started to feel excited at the prospect of learning more.

The two and a half years that I’ve been working my current job have been a learning curve in itself, so I’m trying to treat the Digital Lab in much the same way. I’m still not quite sure what the next step is within the framework of my experiment but adding all of the Digital Lab knowledge to my brain can be nothing but helpful. Whether it’s the fear of failure or fear of the unknown, the best way I know to quieten that fear is saturation via education. (Feeling confident enough to apply that knowledge into a practical solution is, however, a very different question but that’s for a whole other blog post!)

Header Image courtesy of Liverpool-Biennial 2016 © Joel-Chester-Fildes 

Streamlining #DigiLab

Sarah Dick at the Royal Institution talks about streamlining her content plans with the help of the Digital Lab

At the Royal Institution (Ri) we were entering our busiest time of year, with our Christmas Lectures at the centre of a number of fundraising and PR campaigns. Alongside this we want to extend the theme of the Christmas Lectures to our digital audiences by creating content and engaging in conversations on the theme of the lectures, ‘Who am I?’.

In managing our social media I always try to bear in mind the 80/20 rule – that 20% of your content should be about your ‘brand’, and the other 80% should be content that is created and curated to entertain and interest your audience. I roughly divide the content we post in to two categories: marketing content, and science engagement content.

With the uptick in the number of posts around the Christmas Lectures period, it’s tricky to balance our business-as-usual posts with various messages, calls to action and original content, all whilst maintaining our tone of voice and not overwhelming our audiences.

Organisation is key. Last year (my first year on the job) I made a content calendar mega-spreadsheet. Everything went into it, with just a few words describing the content and a colour code to indicate which campaign it belonged to.

It was so bloated and convoluted that by the time I was done feeding in the science engagement content and the marketing content for the multiple campaigns, it only made sense to me in my colour code-dazzled state. It was a little impenetrable to anyone else at the Ri who wanted to quickly check the schedule to make sure their message was getting out. “It’s a simple colour code you see” I told them. “Yes…” they nodded.

A few weeks later and I had lost the context for why I had arbitrarily assigned a certain post to a certain date, and it wasn’t a simple process to quickly re-jig the dates of posts.

Starting the Digital Lab, I knew I wanted one of our experiments to focus on streamlining the various Christmas-time campaigns.

“I said ‘streamlining’ so much that the word lost all meaning…”

Our first phone call with our mentor Daniel was very helpful. We spoke about our backgrounds, what we wanted to get out of the programme and what experiment ideas we had. I think I said ‘streamlining’ so much that the word lost all meaning.

Daniel made lots of suggestions and right after our call emailed us a handful of links to articles and resources, my favourite being this social media content calendar template.

Looking at the way Daniel has structured his template made me realise that in my ‘streamlining’ obsession, I’d been focusing too much on making sure all the messages from our various campaigns, and all the content types, were all in the same place.

“I’d been thinking of my content calendar too much like a scheduling platform (we use Hootsuite), and not giving the content enough space to be fully worked through and developed before locking it in to a calendar date”.

Daniel’s content calendar is structured to make you think about your posts in 2-week themed chunks, which makes it much more practical to manage content production and make adjustments than the 2-month timeline I used last year. The calendar also gives space for you to input the platforms you will share the content on and the influencers and advocates you will contact to amplify the reach of your message. This is especially useful to help get the most out of your content and build relationships. I’m careful of not bombarding our advocates and reciprocal partners with request after request in the run up to Christmas (when their social media schedules are no doubt equally busy!)

With all this extra information for each post I think it will help me keep track of the message, purpose and audience, as well as aiding the actual content production. This structure has also made me realise that at this point in my workflow, the science engagement content and the marketing campaign content don’t necessarily need to be in the same content calendar and I over-fill the spreadsheet by trying to force it. This year, I’m going to split them between two separate content calendars and then streamline the posts by feeding them in to Hootsuite, where you can easily drag and drop posts to reschedule them.

I hope that having a separate marketing campaign content calendar will be clearer and more focused for other teams across the Ri to quickly check to see when their Christmas campaign messages are scheduled.

Finally, the calendars will also be hugely helpful when it comes to analysing how the science engagement content and marketing messages have performed with our audience, as there is space in the calendar to add your trackable URL. That’s the next big experiment we’re working on, something that’s been a long time coming – revitalising our social media analytics dashboards.

Header image credit: Paul Wilkinson, The Royal Institution

Paid Social Media Advertising #DigiLab

Charlotte Angharad of MBD discusses her experiments into paid-for social media advertising on the Digital Lab

Here I am writing my first blog post as part of the Digital Lab. Experiment 1 was to try my hand at paid social media advertising to advertise one of MBD’s events; The Gunpowder Plot.  It’s a large outdoor show and bonfire night event that we hold every year at Boughton House, a beautiful stately home in Northamptonshire.

We have tried various marketing techniques in the past, but haven’t ever used paid social media advertising.  Boughton House is in a rural location near Kettering and is not easy to get to using public transport.  As a result of this, we find that it’s generally a very local audience that visit the event.  I was intrigued to see how social media ads could work for this audience.

As part of the Digital Lab I have been lucky enough to be allocated a fantastic mentor, Devon Smith. Devon has a wealth of knowledge and is a brilliant brain to pick about digital marketing.  The chats with Devon have been really helpful in giving me really practical ‘how to’ information when running my campaigns. She also gave me a really useful ad campaign template that really helped me to get my head round what I was trying to do.

So, down to the campaign nuts and bolts. I decided to run two campaigns simultaneously, one on YouTube (ran by Google ads) and one on Facebook.  Following the advice from Devon I decided to run two video ads in each campaign to see which performed best. The video ads are quite different in style.  One is a typical show trailer format, the kind of video that we have done before. You can see it here.

The other is a more graphic text heavy ad and you can see the difference here.

“YouTube didn’t like my landing page URL as it had the words, ‘Gun’ and, ‘Plot’ in it…”

I found both the Google ads and Facebook ads sites quite confusing to navigate, even with my notes from Devon close at hand. YouTube didn’t like my landing page URL as it had the words Gun and Plot in it. I had to wait with bated breath for it to be reviewed but eventually I got the OK and the campaigns went live!

Both campaigns ran for two weeks. I found both of the platform websites difficult to navigate to see the results.  Once I downloaded the apps on my phone, I found things easier as there was less information to work with.

I scheduled a chat with Devon for day 9 of the campaigns so we could assess the performance of the campaigns so far and make any necessary changes.  As usual, Devon was super helpful and gave me some really good advice.  In both campaigns the show-style trailer was out-performing the graphic text ad so Devon suggested to concentrate my budget on this ad to get the best results.

So, the campaigns are over now and what have I learnt?

With YouTube my ads reached loads of people, with a total of 16,000 views. To get this level of reach the spend was approx £350.

“…the cost is so low you can afford to experiment”

With Facebook the reach was a lot lower at a total of 974, but the cost was tiny at just over £5. The world of social media ads is still a little bit like dark magic, but I definitely feel more able to navigate it, albeit slowly. It is certainly a form of advertising that I will try again to hone my skills further.  I think I will play about with different ad formats on Facebook as the cost is so low you can afford to experiment. Then once I’ve gathered a bit more research about what works I will dip my toe back into YouTube and Google ads.

My advice to anyone would be to give it a go, but try to get yourself a “Devon” to hold your hand through the process.

Now onwards and upwards to Experiment 2; MBD’s Digital Advent Calendar…….


Header image courtesy of AMA conference 2018, Brian-Roberts Images. 

Data-driven decisions for website redevelopment #digilab

Ray Clenshaw and Emma Rowan at Belgrade Theatre, Coventry discuss their Digital Lab experience so far

We were both incredibly excited to be accepted onto the AMA’s Digital Lab. For a long time we’d discussed the increasing importance of making more data-driven decisions and whilst we use Google Analytics and other digital platforms regularly, we felt we were not unlocking their full potential.

In our first meeting with Devon we focused on gaining a better understanding of analytics which would inform our future experiments. Devon was great at helping to demystify some of the terms used and to identify key reports which would provide invaluable data moving forward.

With a greater understanding of what we could learn via analytics we started to think about what we really wanted to know. With a rebrand of our website on the horizon, we were keen to learn more about the user journey on our current site. We were also keen to explore Facebook advertising in more detail, to examine elements of the platform which we weren’t currently using.


Using data to inform website redevelopment

Our website is getting on a little bit now and we’re just starting to formulate our plan to redevelop the site. It’s been incredibly useful to discuss that with Devon and to identify some of the information we can gain from analytics to inform decisions we make going into that process.

The plan is to explore conversions and the user journey behaviour to see how people currently use our site and to gauge what people want most from what we offer. We’ll also be listing assumptions we have as an organisation about the content on the site and how people engage with it, and then finding ways to test those assumptions to see if the data backs them up or not.

Making ‘data driven decisions’ is something we talk about a lot, and we obviously use data from our CRM and ticketing systems all the time, but we’ve always felt there’s so much more we could be taking from google analytics to inform all kinds of activity, and that’s exactly what the Digital Lab experience is allowing us to do now. It’s really exciting and will no doubt prove incredibly valuable as we increase our knowledge and understanding.


Remarketing versus boosted Posts

Facebook advertising is a key component of all of our marketing campaigns but we felt there were elements of Facebook’s advertising platform that we were yet to explore, particularly the remarketing function. Whilst we’d both been exposed to remarketing ads personally we’d never used them as part of our social media advertising strategy. We were keen to examine whether they were more successful than the boosted posts and newsfeed ads we’ve traditionally used.

Devon was really helpful in talking through the logistics of setting up a remarketing campaign and helping us to ascertain what our goals were and how to track them via analytics. Once we’d established our goals we decided to set up an experiment to compare the success rate of a remarketing campaign versus a boosted post. Focusing on one of our upcoming Christmas shows we set up a 2 week remarketing campaign with, ‘number of transactions’ as the trackable goal. Once the advert has run its course we’ll run a boosted post with the same budget and compare the results.

Making changes to tried and tested marketing tactics can be scary, particularly when budgets are tight and the results are unknown. Being part of the Digital Lab has helped to increase my confidence in taking ‘risks’ by improving my knowledge of analytics and how to set goals, track and evaluate the success rate of campaigns. It’s been great to have the opportunity to test some of our ‘what ifs’ and I can’t wait to see what the results show us.

Header image courtesy of People United © Janetka Platun — BELONGING

Changing tact — using film to create rich experiences for existing audiences #DigiLab

Tom White, Assistant Digital Manager at City of Birmingham Symphony Orchestra talks about his early involvement with the Digital Lab

New audiences are pretty much the arts marketer’s holy grail – it’s no secret that getting people’s attention and persuading them to come and experience our art for the first time can be a challenge. But something we’ve been thinking about lately at the CBSO is the importance of keeping that attention, and how we can stop those hard-won attenders from ‘lapsing’ after they’ve taken their first steps. So my first discussions with my AMA Digital Lab mentor focused around the ‘visitor experience’ – not necessarily a concept we deal with too regularly as a performance-based producing organisation – and ways we could enrich our concertgoers’ experience digitally.

Digital video is a format that we are already quite comfortable with as an organisation, so we have started doing some experiments around how we can use it more as a re-engagement tool than a mass-marketing medium.

On any given week, the CBSO will perform between one and three entirely different musical programmes, usually at our Birmingham home venue Symphony Hall but also on UK and international tours. As a result, our production cycle is very tight, with no more than five days between the first rehearsal and the end of the concert – and then maybe only two or three days until rehearsals start for the next one.

As far as generating video content to support concerts online, within that week-long window we can film rehearsals, interview conductors and musicians, and quickly turn around social media videos in time – just about – to publish them on the day, or day before, a concert. We’ve done that quite a few times, using video as a last-minute sales push. Admittedly the content has been very useful in terms of social media engagement and reach – but it’s not necessarily helping us sell seats in the hall. We already know that our core audiences will book several weeks out from the concert, so it’s impossible to get this content in front of would-be attenders quickly enough to make a difference.

So, as part of the Digital Lab, and with our lapsing booker challenges in mind, we’re experimenting with a change of tack. It’s really simple – the plan is to also use video post-concert instead, with the hope of creating a richer experience for existing bookers. A no-brainer, right? Our team already sends post-concert emails after every performance (with press reviews, a social media roundup and recommended concerts), so we were aware of their usefulness. But the Digital Lab has given us the impetus to try video in this context too.

In practice, for a few trial concerts this has meant collecting very short (60 seconds or less) video interviews of musical soloists or conductors immediately after they come offstage – literally the musical equivalent of post-match footballer interview. The clips are then shared directly with the attenders of that particular content via email. Of course, these videos aren’t going to break any YouTube views records – in fact they’re not even visible on our channel – but that isn’t the point. Our hope is that doing this can enrich the audience experience in a way that might increase a first-timer’s engagement in the artform more widely, and make them more receptive to further marketing communications from us in particular.

We’re measuring the responses to the videos, and early results seem positive, with CTR substantially higher on these content-rich emails than on our standard post-concert messaging. The real litmus test, though, will be whether this has any bearing on how those same people respond to the next enewsletters or concert eflyers they get from us. Will the video engagement help to punch through the white noise of marketing in people’s inboxes, or will it make no difference at all? Either way, we’re looking forward to finding out.

Header Image courtesy of Sound and Music © Martin Weiss 

Get the most out of UTM Codes #Digilab

Devon Smith, mentor on our Digital Lab shares insights on UTM Codes    

One of the great benefits of online marketing is the ability to measure the direct impact of your marketing efforts. In order to measure the success of emails, social media, all manner of online ads, and other online marketing efforts, each of those needs to be tagged with a UTM code.

UTM codes are deceptively simple and complex. You can add a UTM code to any link that you control (links that you add to an email, social media post or ad, a digital ad in an online newspaper, an entry in an online community calendar, etc), and the information in that code is then automatically passed to Google Analytics. In short – by adding UTM code to links, you can get more and better data into your Google Analytics, so that you can evaluate whether, for example, visitors from email bought more tickets than visitors from social media, or exactly how many donations were made by visitors who arrived as your site from that end-of-year giving email campaign.

That’s the easy part:

Link + UTM Code = More data in Google Analytics

The hard part is making sure all of the UTM codes on all of your marketing links use the same structure and conventions, because Google Analytics is maddeningly precise in how it organizes those UTM codes in analytics reports.

Before we dive into figuring out how to solve that issue, first a little nerd history: before Google Analytics existed, there was a similar service called Urchin; it was Urchin that developed the “Urchin Tracking Module” code that we now shorthand as UTM. Google acquired Urchin and folded in the Urchin Tracking Module (UTM) code method into Google Analytics. This is why Google automatically attaches UTM codes to all Google AdWords ads by default (and so you shouldn’t be adding UTM codes to your AdWords links!).

UTM code must have the “Source” parameter (a parameter is the type of information being passed to Google Analytics) and can have up to 4 additional parameters (so if you don’t want to use any of the 4 parameters, just leave them out of the code entirely). Each parameter follows the same structure: utm_parameter=youdecidewhatgoeshere and those parameters are joined to each other with an ampersand (&).

Source: intended to identify the name of the website where the link is placed or the name of the email client (particularly if you are a large organization that, for example, uses Mailchimp for some emails and DotMailer for others). So, your Source parameter is often identified as facebook, twitter, mailchimp, theguardian, etc.) and looks like:


Medium: intended to identify the type of marketing channel where this link is being placed, for example: email, social, referral, display, etc. Google Analytics has very strict rules about how it will use the Medium parameter to identify traffic in the default “Channels” report. If you don’t following Google’s rules for the Medium parameter, lots of your traffic will end up identified in the “(Other)” category for this report. After tagging all of your links with UTM codes, you want a Channels report that looks something like this, where the “(Other)” category is quite small:

Default channel grouping chart example

All the rules that Google has for formatting and using the Medium parameter can be found here, but the basic idea is for traffic to end up in these channel groupings. For traffic to appear in the Channel of:

  • Social use Medium of exactly: social, social-network, social-media, sm, social network, or social media.
  • Email use Medium of exactly: email
  • Affiliates use Medium of exactly: affiliate
  • Referral use Medium of exactly: referral
  • Paid Search don’t add UTM codes to any of your Google AdWords! But if you use other paid search network, (Bing, Yahoo, etc) then use Medium of exactly: cpc, ppc, or paidsearch
  • Other Advertising use Medium of exactly: cpv, cpa, cpp, or content-text
  • Display use Medium of exactly: display, cpm, or banner


So your Medium parameter is often identified as social, email, referral, or display and looks like:


If your Medium parameter doesn’t exactly match one of the terms in that list (spelling, capitalization, spacing must all match), then your traffic will get grouped in the “(Other)” category in this Channels report. If that happens, or for some reason you can’t follow the default definitions of Medium that Google has created, you do have the option to update those definitions in Google Analytics!

You can create your own rules for how to sort traffic based on UTM parameters in the Administrative section of Google Analytics, under “Custom Channel Grouping.”

Campaign: intended to identify the name of the thing you are promoting. Ideally, you will have lots of different Sources and Mediums that all share the exact same Campaign code. This is often the name of the performance, exhibit, course, or fundraising campaign. It’s a best practice to use dashes or underscores to separate words, use all lowercase, and minimize the number of words used so your Campaign parameter might look like:


Term: Google uses this parameter specifically for Google AdWords to identify what keyword was bid on, but YOU can use this parameter for anything you want! It’s a total free-for-all. If I’m tagging a social media ad, I’ll often use this parameter to identify the audience segment I was targeting. If I’m tagging an email, I’ll use this parameter to identify the date the email was sent or the list segment the email was sent to. So use the Term parameter for whatever you want that you think will help you better analyze your marketing efforts, and it might look like:


Content: google uses this parameter specifically for Google AdWords to identify which ad the link was placed in, but YOU can use this parameter for anything you want! Again, it’s a total free-for-all. If I’m tagging a social media ad or an email, I’ll often use this parameter to identify the type of content in the social post or email, such as a video, slideshow, or image, or I’ll use it to identify the specific text copy that we might be testing. So use the Content parameter for whatever you want that you think will help you better analyze your marketing efforts, and it might look like:


Put it all together: UTM codes are joined to the link that you’re posting by a question mark and by convention typically appear in the order listed above. So we put all of those things together, and your link might look like:

This is the full link that you would include in your email newsletter on 31 December using a 30 second video to promote Beauty and the Beast. If there are two different places in that same email where you’re promoting that show, just change the Content parameter. If you’re also promoting the show via a video Facebook ad, change the Source, Medium, and Term (but leave Campaign and Content as-is).


Google created the “Campaign URL Builder” website to help you create that long link:

‘Enter the website URL and campaign information example chart’

There’s also a plethora of free or cheap tools you can find online that help automate this process, including Excel add-ons, Chrome extensions, Google Sheets add-ons, etc. Do a little googling to find the one that fits your needs best.


If you decide to go the manual route, it’s still great to use a spreadsheet to keep track of all your parameters throughout a campaign, and that spreadsheet might look like:

Example spreadsheet

Capitalized letters can wreak havoc on your well structured and planned UTM codes. For example, if you accidently use “beauty_beast” and “Beauty_Beast” Google will think that those are two different campaigns, and not group the data together in your Campaigns reports. If you think you might get a little sloppy with capitalization (I know I’m guilty of that!), create a filter in Google Analytics that will force all URL data into lowercase automatically:

Creating a filter in google analytics example

Have you lost sight yet of why we started using UTM codes to begin with? Now that your UTM codes are on all of your digital marketing links, you can start using them to understand which source, medium, term, and content is performing the best for each of your campaigns. So if everything went as planned, you have a nice organized list of campaigns:

Campaign list showing acquisitions

And you can click on any of those campaigns, and see which marketing channel is generating the most ticket sales:

More campaign acquisitions information

Or which audience segment that you created for a Facebook ad campaign is generating the most ticket sales:

example sheet showing which audience segment is generating the most ticket sales:

Or which type of content is generating the most ticket sales:

Once you understand the rules and conventions of how to use UTM codes, there’s a world of analysis that’s possible in Google Analytics – hopefully making more efficient use of your limited marketing budget. But it’s not enough for a single person to follow all the rules and conventions – everyone at your organization has to share the same practices! I often find that’s one of the hardest parts of campaign code management – just keeping everyone on the same page and using the same naming conventions. It’s helpful to share a single spreadsheet that tracks all UTM code parameters at the organization, and to have regular meetings with all of your marketing staff who might need to use UTM codes in their work so that everyone can work out discrepancies.

Good luck with your campaign tagging and may the magic of UTM codes help all of your online marketing efforts succeed!


Header Image courtesy of Liverpool Biennial 2016 © Joel Chester Fildes — Rita McBride, Portal, 2016.  

Change of details?

If you would like to change your contact details or organisation please get in contact with us.