SØLV - Her er en stærk case med rigtig flotte resultater. Det blev på målfoto, at DR Koncerthuset måtte gå hjem med sølv og ikke guld.
Personalisering og omnichannel er lige efter lærebogen, og viser tydeligt, hvad personalisering kan gøre af forskel Vi glæder os til at følge casen, når den bliver implementeret på flere platforme og til at se om indsigterne fra kampagnerne i fremtiden vil bidrage til produktudviklingen og udbuddet i public service.
DR Koncerthuset – concert hall of Copenhagen
DR Koncerthuset is a renowned concert venue in Copenhagen that hosts 400 concerts in rock, pop, jazz and classical as well as family events for around 350.000 guests per year.
DR Koncerthuset has gone a long way providing their customers with relevant digital communication: segmented, automated newsletter, onboarding flow, satisfaction measurements, targeted SoMe campaigns etc.
Measurable objectives for DR Koncerthuset:
DR Koncerthuset felt that there was potential to deliver an even more personalized experience to the customers and leads and thus improve communication performance:
• Increase sales rate and customer engagement
• Create differentiated communication content through automation
• Improve email relevance and reduce spam
• Improve conversion from email, web site, and SoMe campaigns
Add a scientific approach to omnichannel marketing
DR Koncerthuset and ONE Marketing decided to experiment advanced technologies such as Artificial Intelligence (AI) and Machine Learning (ML) in order to increase the communication relevance for each single customer, across all channels.
AI and ML allow to process huge amounts of data and to understand what the customers really want.
DR Koncerthuset and ONE Marketing developed a pilot solution between September 2017 and October 2018 and ran different series of tests in the period.
The solution was adjusted throughout the period and finally ran into production in February 2019.
What does the solution do in essence?
We analyze huge amounts of data to score the customer’s or lead’s interest for a concert and the customer’s likelihood to purchase a ticket for this concert at a point in time. This likelihood is used to pick up the most relevant concert(s) per customer and lead and display it omnichannel: Email, website, SoMe.
What data are used?
We use traditional modelling and response data:
- Traditional data: Socio-demographics, transactions, customer’s registered interests, etc.
- Response data: Website click history such as “read more” clicks, email click history, waiting list registration etc.
What do customers and leads experience?
Customers and leads experience that DR Koncerthuset’s communication content is
relevant for “me”:
- Website: The top banner is personalized [in real time] with the concert that is the most relevant for the customer or lead. The next boxes display the concerts with the 2nd to 7th highest priority.
- Email: display only the most relevant concerts, in prioritized order.
- SoMe: Facebook and Instagram campaigns display concerts that are relevant for the individual.
What can we learn from the data using AI and ML?
Using AI and ML, we can establish a “cognitive profile” per customer or per cookie. We transform the content that the customer or lead has shown an interest for, through clicking “read more” on the website, purchasing a ticket, registering on the waiting list etc., into inputs to modelling.
This gives us a very detailed understanding of each customer and lead, which is individual for each lead and profile.
How is it relevant to increase conversion?
The cognitive profile is matched with the profile of each new concert to assess the relevance of this concert for the single customer or lead, giving a score about how likely it is that the lead or customer will have an interest for this content.
As an example, if you are interested in Pernille Rosendahl which is a Danish Singer/Songwriter, you may as well have a high score for a choir concert (“pigekor”) featuring Kira Skov – even though the two concerts do not address the same audience at first glance.
Is this solution GDPR compliant?
The solution is fully GDPR compliant as customers and leads have given permission to data processing (email permission, cookie permission).
Moreover, all data to perform the analysis is anonymized.
Live test November 2017: +82% more tickets
Six Classic concerts needed a push in ticket sales at the time of test. The generic segment would include 63.000 newsletter subscribers with interest for Classic concerts.
We selected only 11.700 customers with high likelihood of being interested in the set of concerts. Each customer received an individual mail with their highest recommendation highlighted in the subject line. The rest of the recommended concerts were displayed in their ranking order. The test ended with a total of 942 different email combinations, each including 3 to 5 concerts.
The initial results were impressive compared to the generic newsletter:
• Opening rates increased with 14%
• Link tracking increased with 24%
• Ticket sales increased with 82%
Anne Buchwald, Senior CRM Manager - DR Koncerthuset:
”Customers getting Personalized e-mails containing Events that are relevant to them, buy 82% more tickets than Customers getting our normal Newsletter. Our Conversion Rate has increased 5 times after we have used ONE Prediction to select the right audience for our Concerts”.
Live test August 2018: +354% more tickets
Seven concerts needed a sales push at the time of sendout.
22.785 customers in the segment were randomly split in two:
A. 17.102 received a personalized email comprising the 1 to 7 relevant concerts, in prioritized order and with the name of the top one in email title.
That led to 2.275 email variants
B. 5.683 received the “business as usual”, undifferentiated email.
Sales per customer in the segment of customers receiving a personalized communication (segment A) were 354% higher than sales per customer in segment B.
Tech is not enough
Success in marketing automation can never be achieved through technology only. The product and its presentation count as much as the logic behind how it is brought to the customer. DR Koncerhuset maintains a high focus on quality design, portraying a professional touch on showcasing events.
Results in website and SoMe:
As the solution is rolled out to those channels in production in Q2 2019, we do not have results at this point. Expectations are high.
Status February 15th:
The AI and ML based recommendation runs in production and delivers real time scoring.
The pilot solution became a production setup, leading the ground to establishing a cloud service, now available as a stand-alone company: https://www.oneprediction.com/