When Google introduced the Knowledge Graph in 2012, they initiated the shift to semantic search -- the greatest shift in search since the introduction of the Page Rank algorithm. Instead of relying on documents connected by hyperlinks, the Knowledge Graph signaled Google’s move toward supplying answers based on knowledge embodied by entities connected via a graph of relationships.
Digital marketers can learn a lot from Google’s shift to semantic search.
This technology can be applied beyond algorithmic search to improve almost all elements of digital marketing. Brands can create their own graphs to better understand customer needs and improve marketing messages, or they can make sure their solution providers employ graphs that can be easily leveraged.
Taking it one step further, any marketer who doesn’t have a graph within the next twelve months will be at a permanent competitive disadvantage.
What’s So Special About a Graph?
Graphs are fundamentally about two things: entities and relationships.
From a marketer’s perspective, entities in a graph could be publishers in their media buy, queries entered by consumers, ads, promotions, products, orders and customers. These entities would be no different than what one might expect in a relational database.
In addition, graphs contain relationships. In a relational data set, relationships are encoded in the structure of the database itself – all entities in one table have the same type of relationship with all entities in another table.
In entity/graph data sets, the relationships are just as important as the entities themselves, and the direction and the strength of the relationship can also be recorded. Facebook-like relationships such as “is friends with” and “likes” are less important to the marketer, and more important relationships might be: “clicked on,” “was shown,” “bought,” “was bough with,” “was cross-sold,” etc.
Marketing is all about creating and fostering relationships with customers. Moreover, big data is all about finding relationships within large datasets. In a world of both relationship marketing and big data, a graph data set becomes an essential weapon in every digital marketer’s arsenal.
Digital marketers can use graphs to improve their digital marketing in a number of ways. A few examples are listed here:
1. Improve relevance of on-site search engine results
Almost every e-commerce site has a search function. If Google and Microsoft can use semantic search to improve the quality of its search-engine results, marketers should employ the same technology to improve on-site search results.
If the marketer’s graph contains common search terms and queries, those search terms can be mapped to specific products in the marketer’s product catalog, on-site search results can be more relevant and site abandonment rates can decrease.
2. Improve cross-selling effectiveness
Every product page, shopping cart page and post-check-out page can have more appealing “other products you might be interested in” cross-sells.
A graph can improve cross selling with two different relationships. One relationship can be “similar to” products to allow consumers to easily compare before buying. Another relationship can point to accessory products that are commonly “bought with” other products, e.g., a memory card is commonly purchased with a digital camera.
3. Improve SEO traffic
A digital marketer’s primary SEO levers are the organization of their site and the content on product pages (title tag, meta tag, product title, product description). With a graph, both site organization and content on product pages can be changed to better match high volume and/or highly qualified search queries. Better site organization and product pages will then improve SEO placement and traffic/revenue from natural search.
4. Keep paid search campaigns in sync with inventory/product catalog changes
One of the biggest problems in search engine marketing (SEM) is keeping your keyword campaigns in sync with out-of-stock events and perpetually changing product catalog.
Using a graph, you can maintain a link between all search terms, keywords and products. With this link, you will know immediately that, when a certain pair of blue suede shoes goes out of stock, if you should pause the keywords “blue suede shoes,” “blue shoes” and “suede shoes” in your SEM campaign, or if you should leave them on.
Conversely, when new products are added to your campaign, with a comprehensive graph, you should be able to almost automatically generate all the traffic-generating keywords that should be added to your campaign to drive traffic to these new products.
5. Get everyone to speak the same language
As a digital marketer, you sell products (and/or services) to your customers, and those products are organized into different departments and categories.
Yet, especially in search marketing, you hear people talking in the entirely different language of keywords, match types, product targets, ad groups, negatives, campaigns and bids. If you want to compete in paid search, you not only need to know this jargon -- you need to organize, measure performance and manage using this jargon as well.
With a graph, you can keep a map of keywords or queries that map to your products and also use the graph to organize and keep track of all of your search marketing assets. The graph serves as your dictionary or your translation layer, so you and the entire paid search marketing team can talk in the language of your products and not worry about all the search jargon.
Where’s Your Graph?
Creating a graph data set is, like all technical challenges, a build versus buy decision. Technically advanced marketers can choose to develop their own graph data sets. Marketers with fewer technical resources should look to buy applications that have, at their core, a graph data set that is essentially pre-created for them and that they can immediately reap the benefits of.
In other words, to be fully prepared for the semantic search future, digital marketers should be asking all their solution vendors, “Dude, where’s your graph?”
The great shift to semantic search is on. Marketers -- it’s time to get your graph on.
Title image courtesy of Robert Lucian Crusitu (Shutterstock)