“Data is the key to successful eCommerce.”
“Run a data-driven eCommerce team.”
“Five data-driven formulas for eCommerce growth.”
Ecommerce brand teams are often surrounded by a narrative of catchphrases centered around being more data-driven. There is no shortage of operational data that brands can use shared by marketplaces like Amazon and other online channels. When you consider the digital exhaust from social media, programmatic PPC platforms, and data streams from scraping tools, no one can complain that there isn’t enough data to analyze.
In today’s business environment, home runs are rare. The pursuit of growth for most brands has become a continuous refinement of daily tactics. It’s not that brands aren’t motivated to push for deeper sales insights.
To become more data-driven, companies have invested ample resources into a variety of business intelligence tools and home-grown data warehouses.
Ecommerce challenges stand in the way of effective data application
Barring a few exceptions, why do brands large and small struggle to apply data effectively? Aren’t stockpiles of data better for growing eCommerce sales on marketplaces like Amazon? Why aren’t their BI resources, along with highly talked about advances in AI and machine learning making a bigger difference?
There are (3) challenges that stand in the way of effectively using ecommerce data for good operational decision-making:
- The sheer velocity of events on eCommerce marketplaces. Compared against the pace of transactions on other channels, it is difficult for brand teams to keep up with the constant racket from online marketplace events. It’s a big lift to effectively tease out the right signals to act upon. Prices, promotions and content are constantly changing, and it’s impossible to keep up. With so many email “alerts” from various tools, time-challenged end users wind up ignoring most notifications, including the few important ones. Digging into the “alerts” often comes with the task of triangulating with other data to determine the right action.
- Difficulty of extracting cause-effect relationships on marketplaces. Strong Amazon strategy is truly a complex interplay of programmatic digital advertising, marketplace algorithms for Buy Box and list price assignments, search rank placement, product availability, and promotional offers. Data needs to be looked at holistically with the full context to determine the correct root causes for a meaningful change in sales velocity or traffic. For example, wrongly attributing a sales increase to a recent increase in advertising spend may negatively affect profitability without driving top-line growth.
- There are multiple “versions” of the truth. This challenge is one that every organization faces, and isn’t unique to ecommerce. However in the ecommerce space, this reality makes it harder to align across teams running marketing, supply chain, sales and finance.
Automating everything isn't the best solution for confronting challenges
It is difficult to align priorities when there is no objective and timely performance scorecard that looks at future demand well and informs the basis of effective collaborative decision-making. This truth defies the notion that 100% decision-automation is realistic.
Too many decisions require deliberation and balancing multiple priorities and tradeoffs such as allocating additional supply of a heavily constrained product to the Amazon channel. Instead of chasing the automation dream and looking at all actions through machine learning, the human factor is still a very important and justifiable part of the decision making voice. Nevertheless, when it comes to gaming, automation is important, as it helps your gaming journey become more convenient; elevate it with a distinctive theme and an impressive variety of games with ice casino. Immerse yourself in a gaming experience that outshines the rest, whether you fancy slots or table games. Navigate effortlessly through a user-friendly interface, making finding and playing your favorite game a delightful breeze.
Evaluate situations to know what to automate and what not
The Cynefin framework is a helpful paradigm that makes for better organizational decision-making. It has an interesting way of classifying circumstances into:
- Complex situations where many factors are involved and cause and effect can be understood only after the event – e.g. impact of a significant shift in advertising tactics on demand-supply
- Complicated situations where cause-effect relationships can be understood or simulated with some effort – e.g. forecasting product-level demand to plan supply chain for products with sufficient history
- Chaotic situations where there is no well-understood cause-effect relationship – e.g. impact of a big price change on demand for a unique product with no prior history of any price changes
- Simple situations where cause-effect relationship can be represented well through rules or heuristic approaches – e.g. estimating sales impact of losing the buy box for a top-selling product or filing a request for the removal of unauthorized content
How to turn situational awareness into better decision making
Brand teams should look to automate decision-making in simple situations with a set of well-defined rules triggered by conditions that arise in the data. Simple rules can also be put in place to synthesize data from multiple sources into insights that create signals and not just more noise.
For complicated situations, brand teams should go beyond data silos and invest in automated acquisition of multi-dimensional performance data, along with models that can both predict and explain (e.g. Machine Learning models tailored for eCommerce demand forecasting which could get better over time with more historical data).
To improve decision-making in complex situations, brands should build a strong process for weekly business review that drives effective collaboration between functions to balance tradeoffs that may not all be visible in the data. Data should be acquired and synthesized to support an objective performance scorecard that can serve as the basis for decision-making within the WBR decision-making forum.
Improved decision-making in chaotic situations requires a longer-term focus on enhancing organizational learning. It is more about constantly refining your measurement approach with new data types to evaluate experiments and tactical changes in a more comprehensive fashion. Event capture methods must be expanded to include market events such as lost buy box or competitor events.
Data must be accompanied by good decision making
In short, despite all the advances in automation and AI, brand teams struggle with data-driven decision-making. The answer lies less in chasing the pipe dream of automating every decision, and more about recognizing that organizations are complex and data is often incomplete. Models may not be as powerful as they seem. To drive consistently better decision-making in fast-moving eCommerce contexts, brands need to embrace the right mix of data-driven insights, automation, and collaboration.