9 Considerations for Effective Matchmaking
Networking and meeting new suppliers/providers remains the top reason to attend an event. Business owners and service providers want to attract the right buyers and influencers, and exhibiting at events… is the most popular way to enable that. Doubtless that the event organisers have a big responsibility to make the visitor attendance and exhibitor presence on the day of the event as valuable and efficient as possible, and it’s not surprising then that ‘Matchmaking’ is seen as the de rigueur capability expected by both the visitor and exhibitor parties to facilitate the meetings and relationships that matter.
We, at Acrotrend, have worked with many event organisers to build matchmaking capability and believe every event organisation can start with some shape of matchmaking and evolve as they go. Data science and advanced analytics is a key element to produce robust matches, but it doesn’t stop there. The success really depends on what approach you take and how you improve the capability via the triangle of data, analytics and feedback processes.
In our experience, Matchmaking is more likely to be effective and successful when the below key points are considered in the approach:
Collection of right data is pivotal
This might sound pretty obvious, but here is where the make or the break happens. You might be already collecting some data about your visitors during the registration process like the job titles, regions/country, interests, purchase role etc, but is important to ensure that it is collected in a way that can used intelligently by the model. How do you ask multi-choice and subjective questions, and which of them are used for matchmaking needs some thought and structure.
It is also important that you collect the same information from the exhibitors/suppliers as well, and with the same choices of answers, so the matching is easy and accurate. It’s very easy to collect a lot of data but it can get messy when the matchmaking algorithms are trying to use it in purposeful ways. And this is just one type of data – expressed or declared by the participants themselves.
The matchmaking recommendations will be much more powerful if you can deduce interests and needs based on behavioural data – like browsing history and searches on event website and exhibitor directories, clicks (types of links/products) on promotion emails, social mentions and so on. This digital footprint and keyword matching can go a long way in discovering needs and actually affirming the expressed interests as well.
Custom or bespoke, what suits you best
There are plethora of matchmaking and recommendation capability tools and platforms that provide ready-to-use services for your events. Most of these tools can be tailored to some extent to be able to use the data from your registration systems, but might be limited to actually use the behavioural and other data that reside elsewhere. Also the matching algorithms are mostly generic and found wanting in terms of depth and customisation for your event specific business rules and logic
However, if scale and time to market is of essence to you and if you are starting on a blank slate and need basic capabilities, and if your events are more or less similar to one another, then tools are easy to get started on. You might soon need to look for alternative or complimentary solutions in case you want a more involved matchmaking capability that really works.
Choosing the right data science model is the key
One of the limitation with ready-to-use matchmaking products is that the analytics algorithms used are not transparent for you to understand or customise and improve upon. This can be a real problem if you want to test different approaches and adopt one that works. There are mainly 2 types of matchmaking and recommendation engines:
• Content-based – as the name suggests the matching is done based on the product/exhibitor attributes that the buyer/visitor is looking for. E.g. if a visitor is looking for heating products/services for homes in London, then the engine picks up all the exhibitors dealing in heating products and with services in London, rates them based on the proximity based on user history, other product attributes. This approach will work if you are just starting to match on minimum data and if the visitor base is not actively engaged on the products/offerings from exhibitors. If you have rich data-set on usage of products, and other behavioural information, then you might want to use more nuanced matchmaking approach of collaborative-filtering.
• Collaborative filtering – works on co-usage and the similarities of one visitor to other visitors who display related interests/needs. This approach can include similarities and proximity in products/services as well. The output is based on the assumption that two visitors who liked the same products/exhibitor in the past will probably like the same ones now or in the future. This approach is less restrictive than content-based and aids discovery of more and new products/exhibitors with diverse and varied recommendations on matches. At the same time, it can be slightly difficult to explain the matches, without proper data science validation techniques.
As the understanding of how matchmaking algorithms works on your data goes up, and the quality of recommendations they produce are validated, you will eventually want a hybrid of the two and also a way to manage the weightages for both/either. Hybrid ensemble to produce the desired level of robustness in the recommendations is definitely an advance data science skill, and needs a lot more considerations on the configurability, scalability and output interpretability. Building bespoke recommendation engine that fits your business model and rules can be very effective if done rightly.
Coldstart is a tricky challenge
Similar to the choice of analytics algorithm above, coldstart is another challenge that needs proper consideration from a data science and analytics perspective. This also needs a close working with the industry knowledge of the event in question. Cold start problem is essentially when the matchmaking engine doesn’t have enough information to produce the right matches, e.g. first time visitors with no registration information, or new products/exhibitors without a past history for the event.
Collaborative filtering discussed above has this coldstart problem when it comes to absence of product usage history, and hence new products may go unrecommended. Content-based filtering can work for such products. For new users without usage history, pushing content based filtering might work. For anonymous users, generally popular products/exhibitors in certain categories might be the way to go. Or maybe, not doing anything and allowing the users some discovery time, fill in the registration data and create a minimum activity trail might be more efficient than letting random recommendations backfire.
The approach should be definitely based on how the industry of the event works best – eCommerce and retail might be very different take on coldstart than how Solar energy related events work.
Matchmaking is a 2-way process
It’s important to remember that matchmaking implicitly has 2 sides – buyer/visitor and seller/exhibitors! It is as important to provide good buyer matches to suppliers and exhibitors as it is important to help your visitor find and meet the right exhibitors. Here is where the data collection from exhibitor profiles is important. Exhibitors typically don’t go looking for buying companies on the websites etc, so you will not have rich exhibitor behaviour data, therefore making sure that you have good quality responses on exhibitor profile data is crucial.
For the same reason, content-based filtering approach for recommendation should be considered in the ensemble, so you can provide balanced recommendations – not just making sure that the every exhibitor new or re-booker has a fair chance of getting recommended and no one exhibitor is over recommended, but also that the exhibitors get the right kind of buyers/visitors matched to them.
So, capturing the information about buyer persona within the exhibitor registration information should be considered. In fact most ready-to-use matchmaking platforms totally miss the capability of honouring exhibitor preferences and expectations within the matchmaking algorithms.
Making it transparent is as important as making it pervasive
Once the capability to create basic and usable matches is available, they should be made visible to both the exhibitors and the visitors on all the channels they interact and engage. This includes not just the email promotions, but also the event websites, registration systems and badging, mobile apps that you might be using.
Making sure the visitor can see their likely business connections on all the channels and all touch points will ensure the recall factor and motivate them to take the next step to actually booking a meeting or adding it on their event schedule and agenda builder. For the exhibitors as well, making sure that they understand the importance of matchmaking and having it reinforced in the sales collaterals, value added services etc ensures that they fill in their profile information in time and to the most accurate details.
Having said that, when the matches are finally delivered, it is important to be able to understand and explain why a specific match was generated for a given visitor-exhibitor pair. Transparency in terms of which data elements are used for matchmaking and how (you don’t have to give away the IP!) will establish trust in the capability and also help you gather feedback and inputs to make it more robust as you go.
Closed-loop feedback is a must
You need to know how the matches that you provided fared, were they too broad or were tailored to specific user needs most of the time? How does your recommended matches influence the visitor behaviour and exhibitors as well? Do visitors find the matches relevant and useful? How does it relate to improving event experience, increasing rebookings and re-registrations, and boosting revenue?
It is important to agree on the performance and effectiveness KPIs for the matchmaking and recommendation systems you build, so you can actually devise a data collection and measurement strategy. This is best done on the systems and automatically, instead of qualitative surveys and interviews. This might mean bringing A/B testing into the rollout and test strategy, and also enabling websites and other event applications to capture actions on recommendations like click throughs, meetings set etc.
When these measurements are fed back into the matchmaking engine model in a supervised or unsupervised manner, you can improve the match relevance drastically.
It isn’t about just the 2 days of events
The real power of matchmaking can be realised when it is not limited to just the days of the event. Ideally the matchmaking should enable connections and relationships all through the year. If your event brands have associated content websites as well, then integrating the capabilities on both together can be really powerful to nurture 365 days of creating relationships in the digital world and reinforcing that face-2-face on the days of the event.
There is always a next level to achieve
Matchmaking can be truly a competitive differentiator when you evolve it as a strategic capability. As you start with lean matchmaking and gather feedback, visitor and exhibitor data quality from registration and demographic data as well as behavioural and activities becomes paramount, to make the matches more relevant and robust. You can also identify special segments into your database with different, niche and specific matchmaking needs relevant to their industry or roles. For example, matchmaking for special interest groups or women-only networks could be highly niche and useful.
Be aware that matchmaking is not for everyone – some of the regular visitors to your shows, or the ones who prefer a more DIY approach might actually get distracted by the recommendations and might want to turn them off. You might want to make it a preference-based capability, or analytically identify micro-segments within the database to be excluded from matchmaking recommendations to prevent frustrations and dissatisfactions.
On the other side, matchmaking is not just about visitor to exhibitors and products. There is immense value in peer-to-peer networking, and so enabling visitor to visitor matchmaking based on similar interests, expertise and needs can be considered. Not only does it allow for visitor networks, but as an event organiser you get a starting point for a great community build up that is very aligned to your event domains. Likewise, some exhibitors might be looking for business partners with for scale or spread.
When done rightly, matchmaking can create and add value to your attendees and exhibitors and help them build meaningful and genuine connections. In short, there are always many possibilities for each level of matchmaking capability, which only increases as you gain more maturity on the approach and capability.
Acrotrend can support you implement matchmaking successfully. Whether your business is brand new to it or needs to take established processes to the next level, our 80-strong, globally-operating customer analytics consultancy is here to help.
Call us on 0208 123 3208 or email [email protected] to book your free consultation meeting.
Contact us to learn more about how we can help: