For the hottest buzzword in the industry, very little is understood about data. At the recent IBTM World show in Barcelona, there was scarcely a session that didn’t mention its importance. Vice-president of European sales at Cvent Jamie Vaughan went so far as to say: “An event planner without data is just an event planner with an opinion.”
We know the importance of gathering data, understanding our audience and having statistics to show stakeholders. But when people say ‘data is the future’, what exactly do they mean? What does a data-driven future look like?
To get a better picture, I spoke to data aficionado Mark Parsons. A corporate financier-turned data scientist, Mark has worked at big accounting firms like EY and KPMG. In 2016, he launched Events Intelligence to help the exhibition industry build its data science chops.
During our conversation, Mark told me that he sees an entirely new class of exhibition organiser emerging in five years’ time. These organisers, he claims, will be so advanced at data collection and data enrichment that they will be viewed as tech companies when being acquired. By achieving the status of a data or SaaS (software-as-a-service) business, they will earn higher valuations, breaking through the current upper bond that exhibition organisers hold.
So, how will they achieve this?
How well are we using our data?
Organisers are sitting on vast piles of data. Most of it, Mark tells me, is in unstructured, unlinked databases – list upon list of names, email addresses, phone numbers and company names with some job titles thrown in. How much use are we getting out of these caches?
Mark’s answer, predictably, is ‘not enough’. An event is a data gold mine – a repeating opportunity to gather first-party data on connections. A major challenge in data collection today is acquiring high-quality data. Traditionally, we have worked with third-party data because it’s readily available, but rarely does it help us understand how groups are connected. Experts are increasingly promoting ‘connection data’ as a much more valuable commodity – something that’s available in abundance at exhibitions.
Exhibition organisers bring together separate communities. Their value is in how effective they are as a bridge. With richer, more connected sets of data, organisers can accurately map out subcommunities, or clusters, and link them together. A comparison can be drawn to a supermarket. If Lidl analyses its receipt data and finds that customers are buying wine and baby formula together, they might move the products closer together. Exhibitions connect people with other people and products, much like a supermarket.
Say, for example, there is a B2B trade show that venture capitalists are attending in search of startups to invest in. Some of those startups will be exhibiting, but others may simply be present as attendees to check out the competition. With better data, you could identify those startups ahead of time and provide insight to the VCs about who might be an investment target.
Network graphs: a new way to structure data
Mark tells me a trick to making better use of data – ‘network graphs’. These are interconnected networks of data, very different to the lists most of us are familiar with. There are two basic units in a network graph: nodes and edges. A node is a thing – a company, person or product. An edge is a connection between nodes – person A engaged with content B, for example.
These graphs are perfect for the messy stacks of data that most organisers are sitting on. Normally, the most laborious process in data analytics is ‘cleaning’ – sorting data into a consistent format. “With the best will in the world, most [organiser data] will never be transformed into something structured and usable,” Mark explains. Instead, organisers can arrange their data into network graphs and learn quickly who is engaging with who and with what. Communities can be divided into subcommunities, or ‘clusters’, and these clusters can be interlinked according to behaviour.
Some of the things you can achieve with network graph data:
- Cluster nodes based on similarity so people, content and companies that are similar are ‘near’ each other in graph ‘space’.
- Understand the connections between clusters – which clusters are interested in each other (e.g. the venture capitalist + startup example) and what content each cluster will be interested in.
- Infer the interests of nodes you have less data about based on their similarity to nodes you have lots of data about – this is a key part of building good recommendation systems.
Data enrichment – the key step
In order to prepare data to be structured into these easy-to-analyse formats, ‘data enrichment’ is the key. When I asked Mark which was more important, collection of new data or enrichment of existing data, he was emphatic: “Enriching your data is far more important.”
First, data analysts generally start with basic cleaning – removing duplicates and standardising fields (‘software provider’ or ‘s/w provider’), for example. Then, they are ready to enrich their data to add more context.
Mark tells me that Events Intelligence uses Large Language Models (LLMs) like OpenAI’s ChatGPT and Perplexity’s API to scrape the web for information. They ‘boil the ocean’ with language-based AI tools to gather information about companies, what they do, how they describe themselves and who they sell their products to.
In this way, Mark can turn a simple list of exhibitors into clusters of highly relevant companies within industry sub-sectors, based on information from their websites. With data tools, he can find out which people or companies are similar, or as he puts it, ‘close together’. LLMs enable Mark to review and classify thousands of datapoints with far greater speed and attention to detail than a human could.
The bigger and richer the dataset, the more the data analysis software will give back. With enough input, it will tell you where your clusters and subcommunities truly lie.
Bringing in the outliers
At the AEO CEO Forum in 2022, Mark spoke about data, explaining that, “the mind likes order and wants to put things in boxes and categories. “People are either attendees or exhibitors, delegates or sponsors. You’re either in this category or that category. But the real world isn’t like that … When you seek to classify things, the outlier cases are always a problem.” With enriched data, companies can bring those outliers into the fold and make better, more complete use of their communities.
Upskill existing talent
One of the most common responses Mark receives from organisers when singing the praises of data enrichment is: ‘We’d have to bring in a specialist’. According to Mark, this simply isn’t true. Most data scientists worth their salt would grow bored with the relatively simple requirements of an exhibition. Rather than hire an overqualified expert at vast expense, it’s much better, Mark says, to choose analytical minds within your business and train them.
“Typically,” he says, these candidates are “marketers or finance analysts”. He continues: “If you pair these people with a developer with a background in ETL [extract, transform, load – a data integration process], you have a potent combination.”
Reticence around AI and data
AI, like data, is an area that the events industry is behind the curve on. Matthias Tesi Baur, former chairman of the Digital Innovation Committee at UFI, explains: “Just look at the size of AI departments in industries like banking or automotive compared to ours. In many cases, we don’t even have dedicated AI departments and the work is done by good old digital departments.”
We should take cues from other industries, he says. “The best way to make the most out of AI’s potential is to learn from best practices in other industries. The hotel booking or online lead generation platforms such as LinkedIn are good starting points.”
What’s more, we should make hay while the sun shines. “Right now [ChatGPT] is the cheapest it will ever be,” Mark says. “Prices will skyrocket as soon as it is deployed on a mass scale.”
If Mark’s predictions are true, the clock is ticking for exhibition organisers. As more companies jump on the bandwagon and enrich their data, a rift may emerge between the tech-savvy few and the uninitiated. Not only will data-led companies run their events better, they will also breach the ceiling of how much events companies have typically been valued. A window is opening – to move beyond the category of ‘events’ and into the rarified sector of tech. To make this transition, companies need to get data-conscious – and fast.
