Snapshot : Your e-Commerce forecast sets out how much you think you’ll sell online. In this article, we cover why you need it, and when and how you do it. From e-Commerce start-ups to well-established brands, we share where to source data, how to analyse it to predict future sales and who needs to be involve in setting the forecast. We also include lessons from our own first ever e-Commerce forecast.
You need to bring together many different areas of functional expertise to sell online.
Two of the most challenging areas are supply chain and finance. They’re a challenge, because most people in those areas have been trained to prioritise efficiency.
But in e-Commerce, you need effectiveness before you need efficiency.
In fact, e-Commerce is usually the opposite of an efficient sales channel.
Delivering container loads of products to retailers is efficient. Delivery costs are spread across multiple products. The retailer manages the distribution.
Delivering individual orders to customer doorsteps on the other hand, particularly the last mile is not efficient. You’ve got high costs of delivery. And there’s more returns and damages, meaning more work for your customer service team.
So, supply chain and finance people often put up barriers to e-Commerce because they go against their principle of efficiency. One process which helps reduce these barriers is the e-Commerce forecast.
This prediction of your future sales helps both those functions plan ahead. This helps them deliver better efficiency as they can build operational and financial plans against the forecast. It gives them an idea of what’s coming.
The e-Commerce forecast versus the e-Commerce target
Often people talk about the e-Commerce forecast and the e-Commerce target as the same thing, but that’s not quite correct.
At the start of your e-Commerce planning, they are the same. The e-Commerce forecast is how much you expect to you’ll sell, and the e-Commerce target is how much you commit to sell. But expectations and commitments are not the same.
Think about it like a New Year exercise plan. Your forecast might say, I’m going to run four times this week and will do 35Km in total. That’s an expectation.
Your target though would be to do 35Km. That’s a commitment.
But, over time, forecasts and commitments pull apart.
Let’s say your first two runs you only do 10Km. Your new forecast might say your four runs will add up to 20Km for the week, based on doing 5Kms per run. But your target is still 35Km, because that’s what you committed to at the start. To hit target, you need to make up the gap.
The e-Commerce forecast
The same thing happens with forecasts and targets when you start to sell online.
You update the forecast when you have new data. With more data, it gets more accurate because you’ve got a better understanding of what’s happening in the market.
You adjust your operational and financial plans based on the latest forecast.
How many materials to order, how many people you need, how much to spend and when you’ll see a return. These all update as your refresh your e-Commerce forecast.
The forecast is a rolling estimate, not a commitment.
The e-Commerce target
Your target on the other hand is usually a commitment. It’s what you’ve told the business you’re aiming to sell. It’s a performance measure that supports your business case and resource requirements (budget, people, time, materials etc). You report on it to the business, and adjust your marketing plans if you start to fall behind the target.
The e-Commerce target gives a benchmark to measure performance. You use it to motivate and reward teams to achieve their goals. It’s a commitment to deliver results.
You can read more about e-Commerce performance and target measurement in our guide to managing an online store, as the rest of this article focusses on the e-Commerce forecast.
Why is the e-Commerce forecast important?
At the start of your e-Commerce planning, your forecast helps you justify resource requirements. As you start to sell, you use it to manage and adjust those resources.
You use it in your :-
- business case – to secure resources, especially budgets, by showing the return (in sales and profits) on investment.
- marketing plan – to plan how much to spend, and where and when to spend it.
- profit and loss – to track your financial performance, and identify when variances occur.
- business performance dashboard – to track Key Performance Indicators (KPIs) including financial and marketing measures.
- manufacturing plan – to plan how many materials to order and plan resources to produce finished goods.
- people plan – to plan team work schedules, both for manufacturing and service led businesses.
How do you start with an e-Commerce forecast?
In simple terms, you can just guess based on “gut feel”. But with so many business uses and consequences, you normally need to put more rigour into your e-Commerce forecast.
How much rigour depends on how much data you can gather to make a more informed forecast.
The forecast has to cover the next 12 months with sales broken down into months or weeks. In some businesses, you’ll also forecast Year 2, Year 3 etc, but most businesses focus on the next 12 months.
The accuracy of your forecast depends on a number of factors, but it’s almost impossible to be 100% accurate. No-one can predict the future with absolute certainty. There are things you can do to get closer to 100% though. How close you get to 100% often depends on whether you’re :-
- an established player already selling online.
- a new player not yet selling online.
Established online sellers
If you already sell online, you can analyse historic sales data, and look for which factors has the most impact. You can then use these insights to help you forecast future sales.
Continuing trend - linear regression
At the simplest level, you could look at the historic sales trend and assume it’ll continue. For example, let’s say, two months ago, we sold 100 units. Last month, we sold 150 units. So this month we could assume we’ll sell 200 units.
The more data points you have over time, the more confidence you can have in this “continuing trend” forecast. You can use a statistical technique called linear regression which draws a best straight line through your historic data points which you can project into the future. You can use most spreadsheet software to run this sort of statistical analysis.
However, unless you work in a very stable and predictable category, it’s unlikely this sort of e-Commerce forecast will stay accurate for long.
Factors that influence sales - Correlations
Usually, there are many other factors which influence sales.
All affect your e-Commerce forecast accuracy.
You can use these data sets (like advertising spend, website visits, CRM measures) and look for statistical correlations with your sales level.
Do your sales numbers go up when you spend more money on media for example. Do they go down when competitors advertise heavily?
If you can work out for example that a x% price discount delivers a y% increase in e-Commerce sales, you can adjust your forecast to build that in the next time you run an online promotion.
If you have enough budget, you can also carry out econometric modelling to get a more accurate e-Commerce forecast. This is a highly specialised statistical model involving large amounts of data gathering and advanced statistical approaches like cluster analysis. However, it takes time to gather the data and you need specialist to do it. This doesn’t come cheap, so it’s usually only bigger e-Commerce players who can afford this.
While it comes with a high accuracy level, even this approach can’t guarantee 100% accuracy. Customer’s tastes and preferences will change over time. They’re influenced by competitors and innovation. Advertising and media can wear out. Econometric models will struggle to predict these changes.
Manually review of statistical models – Bayesian
Whatever statistical approach you use, you should still manually review the forecast and apply common sense thinking to it. Does it make sense? Do you know of any “new” events that will have an impact. Innovations, new campaigns and competitor activity for example.
Don’t feel wedded to a statistical model if you have better knowledge about the future. You can improve the accuracy of a statistical model with your own knowledge of the market. (This approach comes from an area of maths called Bayesian statistics).
So, for example, think about factors like seasonality. Do you need to adjust the numbers because you know you’ll sell more in the summer than the winter. Or vice versa. What about peak selling times like Easter, Black Friday and Christmas? Your statistical e-Commerce forecast model doesn’t know when those occur, but you do.
What about competitors? If you know how often they advertise or price discount, and when they’re likely to do it next, you can the model to make your e-Commerce forecast more accurate.
New online sellers with no data
Clearly, the more you know about a market, the more accurate your e-Commerce forecast will be.
But what if you don’t know much about the category? You’re launching an innovation, or your just starting out selling online.
Well, the same principle applies. The more you know, the better your forecast. But you have to be more creative in how you source data. And you have to accept that your forecast will start less accurate, but it will become more accurate as you gather more data.
Benchmark versus other categories and channels
If you already see through other channels, you can use this data as base for your online sales. You could use industry standards for e-Commerce selling vs traditional channels and use to estimate your online sales. So for example, the latest NAB online retail sales index puts Australian e-Commerce at 14.3% of the total retail trade. (though bear in mind the e-Commerce boost from Covid-19 and recent lockdowns. See our article on online alcohol sales for example).
It’s unlikely you’ll hit 14% of your current sales in month 1. But assuming you might get there in month 14, a e-Commerce forecast of 1% of your current sales in month 1, 2% in months 2 and so on is not an unreasonable place to start.
Of course, if you use this approach, you need to work out how many of these sales were incremental new sales, where customers wouldn’t have bought versus cannibalised sales, where customers switch from buying in another channel, where they’d have bought anyway.
Online customer data
If you already have a website and you’re adding an online shop to it, you can use your digital data about the visitors you already get. You can make some assumptions about how many of those shoppers are likely to convert to a sale. However, as per our previous article on the D2C business model, this conversion rate is usually low at around 1%-2%.
But obviously, you can use digital media and social media to drive more visitors to the site. More visitors means more sales opportunities and more conversions. You learn how well these work over time, and adjust accordingly.
The reasonable assumptions approach
How much data you can source varies wildly by company and category. But even if you can find out a lot, get comfortable with the fact, that your initial e-Commerce forecast is highly likely to be inaccurate.
Accuracy only comes over time as you gather data and understand the experience of selling your products online. Your e-Commerce forecast gets more accurate as you learn which factors influence sales and which don’t.
You have to make assumptions in the absence of real data, and the best you can do is to make these reasonable. Try to get the rest of the team involved in the forecast, especially the supply and finance teams. The better they understand that it’s a “best estimate based on available information”, they more forgiving they’ll be at initial levels of inaccuracy.
Telling the most reasonable story you can about what you think sales will look like. Be rational, realistic and methodical about your assumptions. But until you have real data, that’s all they are – assumptions. The more everyone understands that and is involved in making those assumptions, the less finger pointing is likely to occur when the forecast isn’t right.
A reasonable approach to e-Commerce forecasting
Using real data helps build believability into your e-Commerce forecast. Data gives people confidence that a forecast is more objective. Even if they don’t fully understand the data, its presence seems to reassure people that you’ve been diligent with the forecast, and it’s not just a wild guess.
In general, it’s usually better to be more conservative in your e-Commerce forecasts. Lower numbers feel more believable and are easier to manage. It’s much better to under- forecast and over-sell.
If you do over-sell, you may have to manage stock shortages in the short-term, but you can use this to your advantage. Stock shortages play on the psychological concept of scarcity (see our article on behavioural science for more on this). They create a perception that a product is popular. This makes it more desirable and can boost future sales.
But if you over-forecast and under-sell, that’s tougher to manage. You end up sitting on lots of stock you can’t sell. You have cash tied up in the warehouse. If you’re products have a shelf life, you risk having to write them off.
Over-forecasting puts pressure on the supply chain team (who need to find warehouse space to store the products) and the finance team (who worry about cash flow and returns). Given they’re already not fans of e-Commerce, you want to avoid this, if possible.
An example from our first e-Commerce forecast
To close off this article, we thought we’d share some lessons from our first ever e-Commerce forecast. (on the same D2C project we cover in our article on barriers to e-Commerce).
Firstly, we had to take on the digital deniers who thought our forecast was too high.
They didn’t believe customers would switch from other channels (mainly because they didn’t shop online themselves).
They laughed when our forecast showed sales growing to $200k/month by Month 12, with an overall year 1 forecast of $1m and Year 2 of $2.5m. We were dreaming, they said.
But then we also had the digital evangelists, senior leaders who urged us to “aim big”, to “go for it”. They told us a modest e-Commerce forecast wouldn’t excite people in the business. We listened obviously, but for us “exciting people in the business” wasn’t the goal.
(see also our article on the three monkeys of e-Commerce as the deniers were Introvert Thinkers – conservative, analytical and sceptical and the evangelists were Extrovert Feelers – energetic, vision focussed and sociable).
In the end, we went with a middle ground e-Commerce forecast. We through of it as a “Goldilocks” number. Not so high that people wouldn’t believe it, but not so low, that it wouldn’t interest people.
Results from that experience
So, what happened in that first year?
Short story was our e-Commerce forecast was horribly inaccurate month by month. We started much slower than expected, but had a change in fortunes mid-year. By the end of the year, we were well above forecast and target. And our Year 2 e-Commerce forecast was way more accurate.
Longer story, was we had a series of unexpected events which threw off our initial forecast.
First, advertising support from the brand team got delayed for the first 3 months. Though we picked up some sales, the sales were modest. Our actual sales were only about 10% of forecast for the first 3 months.
However, the new advertising finally launched in month 3 and we added to the range of products in the store. We used those quiet first three months to improve the store website, the order to delivery system, and help the customer service team get more comfortable dealing with enquiries.
By month 6, we were at 80% of our initial forecast and target. Awareness levels started to rise. Store visits increased rapidly.
The shape of the sales chart was similar to the shape of the product life cycle. Lower at the start but with a definite tipping point where rapid growth took off.
We used this shape in our future e-Commerce forecasts as it was invariably more accurate than predicting straight-line growth.
Sales more or less doubled month on month for the rest of the year.
This took us well over forecast and target. We were inaccurate but had under-forecast and over-sold, and we had a year’s worth of data to make our Year 2 e-Commerce forecast much more accurate.
Conclusion - the e-Commerce forecast
Your e-Commerce forecast sets out how much you expect to sell online over the next 12 months by month or by week.
You use it when you start selling online to validate your business case and your resource requirements. It becomes more accurate over time as you adjust it based on gathering and analysing new data as you sell online.
For businesses already selling online, the more you know they better your forecast.
You can use statistical techniques like linear regression, correlation and cluster analysis to analyse historic sales and marketing data, and make e-Commerce forecasts more accurate. However, you should also apply your own understanding of market dynamics and likely events and trends to improve accuracy. (using a Bayesian statistics approach).
For businesses new to selling online, the same principles apply. But you need to be more creative in how and where you source data. Benchmark against industry standards and other related data (e.g. website visits). But accept that your forecast accuracy will start low and improve over time.
E-Commerce forecasts are a combination of data gathering, logical data analysis, and business storytelling. The forecast is an evolving number that tracks your best estimate of sale based on the latest available data. It helps you manage your marketing, supply and finance plans better by having a view on what’s going to happen.