There are two major categories all data can be placed into when developing your marketing strategy: deterministic and probabilistic data. They both have value, but it’s important to understand the difference between the two and how each can help, or hinder, your goals when working out your campaign.
First, let’s look at probabilistic data. Let’s say Sarah has been reading content on a premium online travel publisher about Bora Bora, but hasn’t yet started searching travel dates, preferred airlines or even hotel accommodation. Sarah may be interested in going to Tahiti one day, but this is an inferred assumption – it is ‘probable’ that she is interested in traveling there, but it’s not guaranteed. Sarah may just be looking for some escapism on a drizzly Monday in the middle of August. Using this information and drawing the dots between Sarah being located in Australia and reading content about French Polynesia would be seen as a probabilistic data attribute.
In contrast, let’s say Sarah searches on a large online travel agency website for return flights from Sydney to Pape’ete via Auckland leaving 1 September and returning 12 September. This is a deterministic data set because it is factual and it can be ‘determined’ without question that Sarah is based in Australia, looking for flights to Tahiti and travelling on specific dates.
Deterministic data will always be more expensive to activate. This is because it is rarer but also because these indicators are a sign of when a consumer is looking at a product in a specific destination and more likely to spend.
Again using Sarah’s example, if a hotel chain in Tahiti buys and activates against the online travel agent’s flight data indicators, that is, Sarah’s search for flights, then it is the optimal time to show the hotel’s brand, price point and offer to Sarah for the greatest chance of making a sale. After all, in Australia 68% of hotel rooms are booked within seven hours of a flight purchase. Activating against probabilistic data in this scenario may end up being a waste of money as it’s impossible to know when Sarah may be looking to travel or even if she is intending to travel at all.
Deterministic data is more expensive than probabilistic data and for good reason. Its return on investment is four times better than the return from probabilistic data, while return on advertiser spend (ROAS) also rises while cost-per-acquisition (CPA) reduces.
Businesses that wish to make a sale now to someone who is ready to buy need access to deterministic data. So next time you plan your campaign, make sure you ask about the type of data set you are being offered because using probabilistic data risks losing out on a sale to a competitor using a deterministic data set.