January 31, 2018 by Ashwini Murthy

Estimating the revenue of an e-commerce company isn’t really difficult. Estimating the eCommerce revenue with a high degree of accuracy where the real difficulty lies. Estimating the revenue with a high degree of accuracy when you have zilch internal data about the e-commerce company available is a Herculean task. At PipeCandy, that’s what we’re doing apart from a hundred other similar tasks.

Being an industry focused predictive marketing platform, we bring the rigor of machine learning models and the depth of data to deliver ‘never before precision’ in estimating online revenue of eCommerce companies. Data is never clean in practice and data sets that point to industry nuances are hardly public. This is where deploying the right analytical models to fill the gaps in data matters. It’s never a simple model. It’s always an ensemble – one to mine texts on pages to understand shipping maturity, another to understand price positioning and a different one to assess omni-channel maturity. The culmination of the ensemble model deployment and unique data sets results in ‘high precision’ estimates of revenue, even for long tail companies that form the majority of global eCommerce landscape.

We’ve already talked about the basics of guessing an e-commerce company’s revenue. The formula is simple:

Sales revenue = Number of transactions x Average transaction value

Average transaction value
Average transaction value, a.k.a., Average Order Value (AOV) = Average product price x Average basket size.

Let’s say you have 5 pieces of clothing in your shopping cart (Basket size), and the average product price is $20, then, the transaction value is 5×20 = $100.

Average product price and Average basket size differ from one category to the other. For example, from a fast fashion brand, an average consumer buys around 3-4 items in one transaction. But for a high-end luxury brand, the average basket size is usually 1 item.

At PipeCandy, we analyzed a few thousand websites to develop the baseline of our eCommerce revenue model. Let’s deconstruct how we went about it (without revealing a lot, that is!)

A simple formula for arriving at the number of transactions in a website will be,

Number of transactions = Conversion rate x Traffic (No. of unique visitors to the website)

Conversion rates are easy to come across (relatively speaking), thanks to benchmark reports released by companies like Monetate. However they lose utility when you start looking at conversion rates at sub-category level (for example, men’s watches convert at different rate than women’s handbags. But most benchmarks would lump them into one category called ‘accessories’).

At PipeCandy, we took average conversion rates for each industry, did sample surveys and baseline data validation exercises with some of our customers and populated our datasets with the help of machine learning models that accounted for more than a couple of dozen variables that had statistically significant impact on conversion rates and basket size. Let’s say the average conversion rate for apparel industry is 1.6%. It’s not going to be the same for apparel companies in all revenue ranges. Given the average conversion rate ‘x’, Conversion rates for top 10% (90-99th percentile) and top 25% (75th to 100th percentile) of the traffic would be very different from the bottom quartile. See the graph below to understand what we mean (the conversion rates plotted are for illustrative purposes only)


Apart from this, we also took multiple other attributes into account which improves the accuracy of our conversion rates, average basket size and average product price. Some of these include:

    • Annual traffic – Accounting for season-based, event-based, trend-based surges and country-based traffic.
    • Product category and subcategory – Example: Category – Apparel & Accessories. Subcategory – Menswear, womenswear.
    • Own brand site/Other retail brand sites/Marketplace. Example: Adidas has its own website, presence on ASOS and Amazon. Traffic and conversion rates on each site might be different. Physical presence contributes to a lift in online conversion.
    • Product quality: Luxury/ Affordable Luxury/ Premium/ Budget/ Discount. Example: Luxury women’s fashion, premium women’s fashion, and Budget women’s fashion all have different average basket sizes, and Average product prices.

There are several other variables at play. Impact of positive and negative reviews and their recency on sales, weather(!), holidays in a month, advertising budgets of an e-commerce company etc. are some more. The list is just vastly bigger but this should give you an idea.


Our eCommerce revenue estimates have a 75-80% accuracy level when looked at as absolute numbers. When looked at as ‘revenue ranges’ they are 99% accurate. How do we know that? We tested.

We checked our eCommerce revenue models with publicly available data and customers’ data. On a sample size of 200 ecommerce companies between $1M and $100M, the average deviation is less than $250K and the standard deviation is less than $1.2M.

Online revenue is the starting point of several segmentation exercises you’d do in a year, as you execute your marketing programs and right segmentation is key to ensure that your marketing dollars are well spent.

With PipeCandy, you can be sure that you have the best data in the market when it comes to eCommerce revenue.

Here’s a quick glossary of all the terms:

  • Number of orders – Calculated based on traffic data and conversion rate
    • Number of orders = monthly visits * conversion rate
  • Average Order Value (AOV)  – Estimated based on product category and subcategories
  • Online revenue  –  Estimated based on traffic data, product category and conversion rates
    • Online revenue = Monthly visits * Conversion rate * AOV
  • Average Basket size – Number of items sold in a single purchase

Want to know more about PipeCandy’s predictive analytical models that power account based marketing campaigns of top tech product companies? Talk to us.

Ashwini Murthy

Content marketer @ PipeCandy

A writer by day. Illustrator by night. Currently trying to conquer the B2B marketing world one baby step at a time. Loves everything outside her comfort zone.