How to battle against Google & Co.?

The value creation funnel in classical retail is interrupted, transformed, power redistributed. Manufacturers and retailers lost a lot of their access to the end consumer. Value constellation in this domain is characterized by the co-petition of several actors and stakeholders, such as eMarketplaces, retailers, search engines, comparison shopping engines, online advertising agencies, technology providers (for retailers and marketplaces), marketing agencies, marketing research, and product data vendors. Extremely fast ecommerce is setting the pace of development and influence the consumer behavior.

We need to accept the new market structure with online marketplaces, search platforms, social networks and other e-commerce platforms own the user attention and influence their decisions on every step in the cross-channel purchase funnel. The speed of data creation is enormous.

Obviously, majority of this data is gathered by big corporations like Google or Facebook but on the other hand essential for successful and profitable existence for retailers, manufacturers, ecommerce players and advertising agencies.

How can SMEs compete against the giants?

In the project EW-Shopp we realized that there is a distinctive group of market players that are creating great customer value and consequently gather ultimate purchase behavior data.

Many of the companies operating in this domain have developed highly specialized competences in integrating and vertically analyzing data in their specific sector. However, this approach produces limited insights and the ability to utilize powerful purchase behavior data remains in the hands of the big tech companies that established platforms for managing cross-domain data infrastructures which are capable of large-scale computing.

The EW-Shopp project overcomes the limitations of specialized domain knowledge by incorporating big data analysis, big data storage and big data visualization specialists into the consortium. Together we can build competitive user behavior insights and demand knowledge that could help European small to large companies to be more competitive and at least maintain the local market position by having the opportunities to innovate their services. Furthermore, the ability to pool market information gathered across such platforms and leverage them through sophisticated analytical tools can offer a wider insight and enable application of lessons learned from one market to others.

Performance Insights: a new analytical tool for vendors and retailers

At Ceneje d.o.o. we have addressed the above problems by providing a tool for vendors and retailers called Performance Insights and an info widget for the end consumers. We have enhanced the simple predictive analysis by considering the weather prediction data and price / stock events on products. As the result the vendor or retailer can foresee changes in category demand based on following parameters:

  • A very strong trend in sales data is present.
  • A seasonality spike is coming in following weeks.
  • Weather forecast will drive the demand.
  • A strong market player has significantly changed prices for a popular product.
  • A popular product is running out of stock.
  • A price elasticity and demand changes based on intrinsic competitive price changes.

The tool is structured in a way to show different dimensions of business performance focused to product assortment management, price development, store relationship, and brand management.

Ceneje is a price comparison platform, meaning that we get product prices from more than 500 retailers and several millions offers daily. This data needs to be organized, cleaned, categorized and merged to have a value for the end consumer. One of the most painful issues is linking together products from different retailers and putting them into the right categories. For years we had a quite complex in-house solution that enabled us to automatically link together and categorize about 60% of the products – the rest had to be done manually. The participation in the EW-Shopp project connected us to Machine Learning experts at UNIMIB and they proposed a new solution to us that is currently at a 90% success rate and promising even better results with fine tuning. Furthermore, the solution is language agnostic, which makes it easily applicable to other countries.

Overcoming the data inconsistency eventually means better user experience for users searching our platform. By doing that they are creating in real time a whole new set of data – the product demand per SKU in hundreds of categories. This dataset is really the core asset for building intelligent sales and marketing business predictions we are adding to our Performance Insights tool.

Through EW-Shopp we gained access to data experts and when we observed the demand data, we concluded that even though the data is seasonal for certain categories it is extremely hard to precisely forecast the spikes. We found out that weather can have a strong impact on sales. Also, marketing campaigns will affect sales. Finding the right parameters and correctly weighting their impact proved to be a several month job for both teams – at Ceneje and JSI.

Initial Performance Insights betas were full of data, tables and graphs. And it all looked great. Unfortunately, it was super useless. We found that out the hard way when presenting it to some ecommerce players in the region and getting back their feedback. We realized that we were sitting on a huge pile of data and that we have the insights, but we don’t know how to present those insights to our partners yet. The solution was to gradually drill deeper and deeper into data and eventually we presented our partners with:

  • The executive view: Get the general idea with a single glimpse on the visuals. How do I perform and compare to market? Are there any forecasted spikes?
  • The management view: React to the insights gained in the Executive view. Focus on certain categories, price, channels or campaigns and develop operational improvement plans.
  • The operational view: Knowing on which categories, retailers or competitors to focus, drill down to the SKU level to make the campaigns and sales initiatives more effective. React on short term market deviations timely and efficiently.

This approach made it possible to get great feedback from different regional players in particular categories to the extent that just recently made a first operational short-term sales and marketing plan with major TV brand producer to boost their market position in one country. Another vendor pushed a defiant retailer to finally become our partner and in this way the market is consolidating and getting stronger and harder for the Google and the likes to penetrate.


Interest in this solutions? Get in touch with us!

David Creslovnik 
Darko Dujic

2 responses to “How to battle against Google & Co.?”

Leave a Reply

Your email address will not be published. Required fields are marked *