Comment: Drowning in data, lacking in insight

Last month I discussed how uncontrolled sales growth can negatively impact online profitability. This month's article is going to explore the problems online retailers face being inundated with a massive amount of data but without actionable insight, and how this dynamic contributes to poor investment and operational decisions.

It has been said many times that eCommerce is all about the numbers and that you can measure the impact of any changes to the customer-facing proposition almost instantly. One of the key advantages digital marketing has over traditional marketing is the ability to measure the impact of digital ads more accurately and quickly than other media; or at least you can if it is properly tracked and reported.

However, this assertion is missing the key point that eCommerce is not actually all about numbers; it is about the insight you can gain from the interpretation of those numbers. Many retailers do not understand the fundamental concept that data without insight is like toast without jam; dry and difficult to digest.

Most retailers are overflowing with data that is expensively processed, correlated, formatted, stored and reported – and then ignored. I recently worked with a large multichannel retailer whose daily sales report was 37 pages long and actually didn't show sales until page three. Given the cumbersome amount of detail it is not surprising that no-one ever read or used the report, and yet it was churned out daily.

Retailers can help mitigate these problems by ensuring reporting information passes the "so what?" test. If you can't answer the "so what does that mean?" question and understand what action to take based on the available metrics, you are asking the wrong questions and reporting the wrong elements.

Understanding that data definitions are the foundations of insight

Quite often, problems begin with the lack of consistent data definitions, causing different systems to produce different numbers from the same data. We have even seen sales within the same company calculated differently in each European country they operated in, making meaningful comparison virtually impossible. You would think sales reporting is easy, but there are a number of ambiguities. Are they including VAT or not? Do returns get deducted automatically? What about cancellations? Is it based on orders placed on the site or when goods are shipped from the distribution centre? Are delivery charges included? You get the idea.

Data definitions also need a consistent time element. Conversion rates and unique visitor numbers are of little value unless you know exactly how they are calculated and over what period of time. Do all systems start each day at exactly the same time? Probably not.

Retailers should never consider implementing any new systems without establishing clear data definitions and should invest the time and budget necessary to understand and document all the key data elements.

Extracting meaning from online analytics

As well as inconsistent input data, the situation is often aggravated by the selection and subsequent implementation of an online analytics tool or, in the worst cases, multiple analytics tools. Additionally, insufficient or improper training can result in data analysts trying to manipulate data with systems that they don’t really understand.

Reports are often based on system capability rather than business requirements. Data analysts all too often simply become data collators, with any real analysis being conducted in spreadsheets by the commercial trading teams who don’t have the time or skills to do it accurately.

In conjunction with these issues, much of the data around site analytics and performance and marketing effectiveness is incomplete, inaccurate or simply does not exist. There may also be insufficient knowledge within the business to challenge initial findings. In many cases retailers measure the wrong metrics altogether and make decisions based upon flawed data.

It is better initially to measure a small number of key performance indicators (KPIs) accurately than to attempt a broader scope. Once the basics are in place you can add more sophistication.

Analysing your true marketing efficiency

The relationship between marketing spend and the revenue it generates is often poorly tracked and misunderstood. Sales are frequently attributed to the wrong channel and there is little focus on true incrementality or customer journeys. In many cases consumers interact with marketing channels on the way to buying products that they would have purchased anyway. In this instance the marketing spend is actually wasteful, but as it appears to generate sales it may receive unnecessary additional funding.

Retailers should ensure that all marketing spend is subjected to a thorough and accurate ROI assessment at the micro level, and activities are fine tuned on an on-going basis to deliver an acceptable ROI. Measuring marketing ROI only at the macro or average level allows poor performing activities to be hidden by the more effective marketing components.

Building an accurate eCommerce P&L

Costs are regularly spread across multiple P&Ls including online, marketing, stores, IT, finance and logistics, but there is no overall accurate view of the full operation’s true cost. Seemingly successful online businesses are often not nearly as profitable as they believe, once all the costs are understood and allocated correctly.

It is important that retailers forensically review the components of their eCommerce P&L to ensure it accurately reflects all the costs of operating online and across multiple channels.

Appreciating that poor insight creates operational problems

The impacts of poor insight across all elements of an online operation often reveal themselves via unforeseen consequences. For example:

1.) Marketing promotes fast-selling lines on the website but not the overstocked lines in the warehouse, leading to poor availability of the best sellers and exacerbating the overstock problem.

2.) Heavy markdowns are now needed to clear the overstocks, which will cause site conversion rates to increase. While this improvement is seen as positive, it is in fact negative; the ‘improvement’ is entirely artificial.

3.) Improving conversion rates often leads to a decision to increase marketing spend. Yet, when conversion inevitably drops, once the reduced stock is cleared, marketing spend remains at the higher level, often with a significant negative ROI.

It is crucial the retailers understand the key interdependencies, as detailed above, in their eCommerce operations. The insight must be readily available to spot any improvement opportunities and to quickly address issues as they develop, and before they have a major financial impact.

Starting in the right place

In much the same way that a golfer cannot fix his own swing without a coach instructing him, you cannot improve or even understand profitability in an eCommerce business without the right high-quality data to gain real and accurate insight into business performance.

Too many retailers start by assessing what data they have rather than what business performance metric they are seeking to improve; this is a recipe for almost certain failure.

Instead of starting with "what data is available?" and then reporting on it ad nauseam, the key question is "what am I trying to improve?" Only once you understand that can you start to define what data is needed to answer the key business questions to help you improve profitability.

Former Carpetright IT director Ian Woosey is a senior director at professional services firm Alvarez & Marsal. He will be writing a monthly column on the challenges of modern retail, exclusively for Essential Retail.

In his article next month, Ian will discuss the significant supply chain challenges retailers face in today's ultra-competitive multichannel world.

Click below for more information:

Alvarez & Marsal