In my last article (in Dutch) I explained how to connect data sources and use data to become predictive. An introduction to data-driven marketing. In this article the subject is explained more detailed.
We received a request of a travel agency to help them be as relevant as possible to their customers using data and data-driven marketing. “Travel agency X: Which steps do we need to take to implement data-driven marketing in our online marketing actions and by doing so, become more relevant to our customer?” So, the job lies with us to advise and provide in solutions and implementation. To get there, we first need to understand the agency's customer. What are the different customer types and what does the customer buying process look like?
These are the basic steps to take first when the wish is there to become more data-driven.
1. Customer types and customer buying process
In data-driven marketing we recognize a few general customer types. The combination of customer types and the buying process stages are brought together in a customer profile that forms the base of your marketing activities. Customer profiles are necessary to segment your customer in a similar group with specific characteristics. Not going into details too much, we use the following customer types:
Gathers information online, is not in a haste to buy.
Did some research and has a specific intention to buy although still showing low activity.
Impulsive buyer, is not looking to buy a specific product or service.
Always hunting for the best deals. Frequent buyer.
Most loyal customer. Has a real connection to the brand. Buys regular and has an above average lifetime value.
The purchase stages combines with the customer types and the organization. Ask the questions, roughly; what is the customer journey of our customer when he or she is looking to travel? What is our customer looking for, is he an Explorer or Passionate? How do you create the Believer?
Customer buying process
Eventually every customer or visitor can be placed in one of the below stages of the buying process.
After recognizing a certain need, the customer will start looking for information about a solution to this “problem”. During a consumer’s decision-making process and information gathering, he or she will pay more attention to what is already known about a product and the information that is being received from friends, family or other consumers. This because this type of information is possibly being perceived as more objective.
After evaluating the different alternatives and solutions for a consumer’s problem, he or she chooses a product or brand that meets these needs best and will possibly purchase this one.
The purchase is depending on the previous steps including the perceived value of a product and a brand, however, it might also be affected by a shopping experience or quality, available promotion and discounts, etc.
After purchasing a product, the customer will evaluate the whole process, including the product itself. He or she will feel delighted when expectations are met, or disappointed if the product fails to meet these. When evaluations are positive, stages of information search and evaluation of alternatives will be minimized the next time and the consumer will possibly be more inclined to buy with the same brand. This likely results in customer loyalty. This post-purchase evaluation might therefore be important regarding the future consequences it has for a brand. Consumers that have a positive experience will likely also share this opinion through word of mouth or other communication channels.
2. The data
Our travel agency owns a few internal data sources with customer information, such as an email database, CRM and booking system. These data sources are connected to the clients’ profile based on a customer ID.
If the customer visits our website, we collect some basic data. Type of device, time and the weather at that specific moment. That basic data is linked to the customer profile.
External data sources
Based on the internal data we have a pretty good idea of who our customer/visitor is and what their interests are. To get really relevant, we need to add external data to this. There are, of course, some very straight forward external data sources for the travel industry. I.e. holiday data, payment of holiday allowance or tax money, weather, and so on. There are also some sources that are not that obvious but much more interesting.
For example, the Dutch Bureau of Statistics, the CBS, provides an API with data about the sport intensity per postal code. This data could be very valuable, especially when our travel agency distinguishes itself by offering active holidays. If the customer or visitor is from an area where the intensity is lower it seem logical to offer other types of vacation. Potential customers that do like the active holidays but live in an area with low intensity will eventually filter out of the data by gathering the click and search behaviour on the website.
Besides the sport intensity, it can be interesting to connect the traffic intensity data API. In addition to the holiday versus traffic ratio (an obvious combination) we can also go a step further with this data. I.e. push notifications with an alternate route to the airport because of a major traffic jam (based on our gathered data we know when and where our customer is travelling). If the jam is extreme, we could also use the data to push notifications with possible places the stay overnight (of course offered through our own tour operator or travel agency).
What influence do social trends have?
A third interesting data source are social trends, or trends on social media. Not the specific brand awareness, but more major events that are occurring. For instance earthquakes, terrorist attacks or other happenings that influence tourism on a specific location. At the location where the event takes place tourism will drop drastically. On the other hand there will be a significant rise of tourism at another destination. If we get this data insight clear, maybe even detect a pattern in it, we can than guide our travel agency in their next steps to jump in with suiting offers. Of course there are much more interesting possibilities that seem less obvious. But combining these data sources above, we can take a huge step towards data-driven marketing actions!
3. The actions
As soon as the visitor enters our site we start collecting the basic data ( device, time, weather). We identify if the visitor is an existing customer or a new visitor and place him or her in the correct customer type.
If the visitor is an existing customer, we have a whole collection of data. We know to what customer type he or she belongs and in what buying phase he or she is in and what their interests are. We already explained five custom types and three buying stages. This means we have 15 (3*5) possible basic “personalized” website/shop publications available. This basic mathematical trick forms the base of the dynamic content on the website/shop. The publication possibilities will decrease every time you gather more insights on your visitor and apply trics (algorithms) to predict the most likely actions of your visitor based on the data.
What if the visitor is new? We do not have a customer profile yet. What do you do?! First, we gather the basic data again. The next step is to place the visitor in the most suiting customer type and while you do this, you keep gathering the click and search data, so your profile is getting more clearer. The most suiting customer type is identified by comparing to the look-a-likes. Which characters are reflected and to what customer type do they belong?
So, now we know how to approach our new visitor as well.
Dynamic website content
As digital marketers we are all well known with the term dynamic content. When we implement dynamic content “data-driven style”, we use the data of our existing customer. We already had 15 possible publications. If we also want to perform the art of persuasive marketing, we add Cialdini’s principles of influence to our profiles. For example the Explorer intents to buy after Social Proof, especially when friend X just bought a trip to Dublin, but the Believer shows more interests when he sees sign of Sympathy.
Cialdini distinguishes 6 persuasive principles. Every principle demands a different customer approach to fit into the explained customer types. So, if we add these 6 principles to our little mathematical trick, we can provide 90 (3*5*6) different dynamic content combinations. A huge step to more relevant marketing!
In short, simple steps: The visitor is recognized -> labels based on algorithms (trics) are added to the customer -> the dynamic content is personalized.
Predictive data-driven marketing
When we use predictive data-driven marketing it works very similar (persuasive notifications are, in fact, mini Next Best Actions). The difference is that the predictive marketing campaigns are much larger and focused on calculated Next Best Actions rather than persuasion. The customer profile, customer type and the purchase behavior (internal data) are enriched with external data specifically fitting to the agencies wishes. In our case the travel agency. The composed algorithm calculates detailed and relevant marketing actions and offers for the customer.
If you want to be predictive in your action, the steps are very similar to dynamic content implementations. Your goal is to steer towards Next Best Actions based on data en customer types. Eventually you implement it in the dynamic content of your website, webshop, newsletter or personal push-notifications, ads and specific offers.
4. Best practice: A data-driven implementation of the travel brochure
Especially in the tourism industry online publications are used in large numbers. Not just as a simple selling tool, but are just as crucial when you wish to gather customer data on your visitors behaviour and interests. In this example we host the digital brochure of travel agency X in a separate environment and compose it there. We add tags to our content to keep up with the click and reading behaviour of our visitor.
The gathered data is saved in the customer profile and helps predict the Next Best Action and determine the persuasive notifications we explained in this article. Of course we combine these with relevant external data sources like the sports intensity per postcode area. Besides that, the content of the digital brochure is easy to adjust to dynamic content. So, all in all, the travel brochure is a very good subject to use in personalizing marketing actions.
Summarizing, the above shows that you can give a very clear direction to your marketing actions of the company and buying process of the visitor/customer. Implementing this, travel agency X now has the knowledge to be relevant to its customers in order to grow conversion rates.
Enough reasons to apply data-driven marketing in your organization (not just the tourism industry of course ;-) )!