The year is 2006. Microsoft is developing the Zune, their attempt to rival Apple’s iPod and take over the market. They do heavy market research and analyze user data to create a superior product. Seemingly, Microsoft had everything they needed to beat Apple’s iPod.
But Zune failed. And failed hard.
Why?
The company relied on bad, misleading consumer data.
For example, during the product development, their market research indicated that users disliked the idea of sharing music wirelessly. So Microsoft, driven by false market insights, decided this feature wasn’t a priority for their product and released Zune without it.
Shortly after, the iPod Touch came out, featuring wireless sharing capabilities. Fast forward to today – we all remember the iPod Touch, but no one remembers Microsoft’s Zune.
Key takeaway?
Bad or misinterpreted data can be worse than having no data when making critical business decisions. For ecommerce retailers, one such bad data point is hidden within brand text search.
Big Problem With Brand Keyword Search in Ecommerce
Most retailers examine their user behavior data so they can use it to improve their platform and make marketing campaigns more effective. However, specific insights create a false image of user intent and can badly reduce search relevance.
For example, your data may show that over 50% of text-based searches rely on brand keyword search attempts. So, let’s say you own a sports clothing ecomm business and that half of the text searches by your shoppers are brand searches, meaning they include brand keywords like Nike and Adidas.
But do all these shoppers really want to buy only the brand they searched for?
Of course not.
They use these well-known brands to search for a style they’ll love. If they come across a pair of sneakers that satisfy their style but are not made by the brand they used in their search attempt – they will surely consider buying those.
And the problem here is that the consumers are used to crappy text search engines that are not good at recognizing their intent and preferred style, so they must resort to a brand keyword search attempt to get the product of a specific type.
This leads to a “false positive” scenario for e-commerce search relevance and a bad user behavior data point implying that they mostly look for brands. While the actual truth is that they know they can’t do it any other way, so they opt for this user journey.
3 Key Damages Bad Data Can Make
Relying on misleading data in e-commerce can lead to several negative outcomes for retailers:
- Misguided Marketing Strategies – Retailers might spend too much on promoting only certain brands, missing chances to showcase diverse items that could also interest consumers.
- Limited Product Diversity – Focusing only on popular brands could mean missing out on unique products from smaller or newer brands, reducing options for customers, and making the store’s selection seem less diverse.
- Ineffective Inventory Management– Relying only on popular brands for stocking decisions can lead to having too many items that don’t sell in stock and not enough of those that do. This imbalance can cause products to sit unsold or run out, hurting sales and profits.
By relying on bad user-behavior data about brand keyword search as an indicator of consumer preference, retailers risk making decisions that don’t accurately reflect consumer behavior or needs.
This can negatively impact their competitiveness and overall business performance.
How to Solve the False Positives Problem in Brand Keyword Search
The research data gathered by the Miros team shows that this can be solved through visual AI search methods and tools.
Miros offers solutions like AI search for ecommerce that can help online retailers through:
1. Wordless Search
With a product discovery tool that understands what’s in a product image, ecommerce platforms can give the shopper an entire wall of relevant products, without relying on a brand keyword search.
The shopper can input one (or none) text-based search, click on a product they like, and then, through a couple of more clicks and scrolls, get a truly personalized feed of products that fit their style.
2. Miros Semantic Search
Semantic search relies on AI to read the product photo and assign countless keywords no human could think of to every single item on an ecommerce website.
This allows shoppers to type in words and phrases, literally any words they can think of, to explain the style they’re looking for. And our visual AI search tool will fetch them exactly what they need.
The main problem with text search is that when a shopper comes to your website, they have a picture in their head of a product they desire, but it can be tough to express their style with words. And this is roughly 80% of your visitors – precisely who Miros is made for.
Here, you can see how it works:
This visual AI search tool acts as a fashion search engine that allows visitors a shorter and more enjoyable shopping journey with minimized text search attempts.
It removes the hassle with website cookies and metadata, while boosting your KPIs and increasing revenue.
Want to help your e-commerce and learn more about Miros? Book a quick demo here.
