9DECEMBER 2023Features like targeting and "lookalike" within these platforms started to be heavily used. Each company had to have accounts on all these data hubs and marketers were only promoting "channel" strategies. Multichannel attribution modelling was (and still is in other shapes) key to understand how to allocate budget and optimise marketing spend. However, in the last 5 years, third party cookies are dying and not even Google is able to give the information we were used to have.What to do now?Everyone is talking about first-party data strategy, and this is the way to go. No doubts about that. You need to collect as much data as possible from the users e.g. email, postcodes, mobile phones. Lead generation is key, optimising your funnel to collect data as early as possible in the acquisition journey. With the smart use of ML you also need to augment that data. From a data perspective this is personally what I consider my checklist for the next 12 months:- Review all acquisition journeys making sure you get first party data as early as possible. Offer content, services, freebies to establish a connection and gather (legally) email and mobile phones- Keep the journey on the web (any device) unless it threatens considerably the User Experience. Jumps between mobile apps are tricky to track- Configure your data pipelines to get a near real-time view of all your leads and customers, making sure first-party data and cookies are connected. This is key to make event-based inferences with any type of model and feed the information back to relevant systems. Example: A lead that visited the website x times, had certain interactions, and downloaded a certain content C can be scored as "high value" from an ML model running on acquired leads. The score is then pushed to personalisation tools or CRM systems to trigger communications and experiences- Enhanced retargeting: pass augmented data (e.g., propensity or value prediction scores) along with first-party data to increase publishers' capabilities and optimise bidding- Set up MLOps to have an efficient and governed way to apply your ML models and Data though is not everything and coming from a person doing the job I do this is a bold statement.In marketing (I might there to say in everything) you cannot expect data to tell you everything. Data will tell you a story of something that already happened, it can give hints, but it will never tell you about things that haven't been tried yet.In the last 10 years marketers and board members have been spoiled, by the large amount of information available. While we are entering a new era where data is becoming (again) hard to find, we need to be brave again and try strategies that could be hard to measure.We need to accept again that for some things we will have error margins, fuzzy correlations, and hints. In conclusion combining a data-driven strategy focused on data augmentation and lead generation, along with a brave and strategic mindset is to me the key to the future of marketing.As the Romans used to say "in medio stat virtus", data will always support business and processes but don't let data block the business. We need to accept again that for some things we will have error margins, fuzzy correlations, and hints
< Page 8 | Page 10 >