Defining Sales Metrics — What Google Searches Won’t Tell You

It should be easy to define how much we’re selling. Selling is so fundamental to business that it seems heretical to even question the concept. Surely this has all been worked out already? We’re all smart people, hasn’t this been dealt with?

Well yes, once upon a time it was. But then maybe the company went multinational, and then global; other companies were acquired, ownerships were transferred, and the definitions that were initially so simple became complex and locked up in layers of data processing code, to the point that the sales numbers are ‘what the database says they are’. Unpicking that code rarely fails to bring to light surprising assumptions and shortcuts.

So why is it so difficult? Search for ‘defining sales metrics’ and you will get any number of results like ‘The 15 Sales metrics you must use!’. The first result I got from that search lists 120 different metrics for sales — all potentially useful, but it doesn’t even begin to help me with an actual, usable definition for the very first, ‘total revenue’. It’s that kind of fundamental definition I am interested in here. There’s much, much more complexity than you probably realise.

Different teams need different metrics

Fundamental to defining a sales metric is the recognition that there may be a need for multiple, valid sales metrics that do not give the same results.

As an example, let’s say I get a marketing email from a company and I go to their website and buy a gift card. Somewhere, a CRM team is celebrating a successful sale, but that sale doesn’t appear on the finance team’s numbers; the money I spent is immediately offset against a liability to send me some goods for no additional money at some as-yet undetermined point in the future. The commercial and operations departments are also probably more interested when they have to supply me with a physical product than in my receiving an email with a gift card code in it.

When I do decide to place my order, the commercial team will put that value against the date I placed the order for that physical product because that’s when they were running their promotions. The finance people are interested in the point that the liability is settled — usually the date at which the order ships.

For many years the holy grail of BI was the ‘single version of the truth’, and that is still a valuable concept — when we look at business numbers, there should be a default enterprise definition that applies (unless otherwise clarified). But when that phrase was first created, the key word that launched thousands of BI initiatives was ‘single’; now we are in the second or third phase of business intelligence maturity, it is ‘version’ that feels more relevant.

Worked example

Here’s a simple example of how two teams can see the same sequence of events as occurring in different amounts, with different people, and at different times:

Scenario

• Customer A buys a £30 gift certificate for their partner on March 31st.
• The partner (Customer B) uses the gift certificate to buy a photobook from the website on April 30th. The photobook costs £27 and there is a £5 delivery charge. They pay the £2 extra with a credit card

Analysis

The CRM team and Commercial team have very different views of this sale:

Why is this? It’s because different aspects of the event are relevant to different teams.

So, the commercial team recognises the sale (excluding tax and shipping) to customer B when they place their order, whereas the CRM team records the money spent by Customer A and Customer B as each makes their transactions (including tax and shipping).

Faced with multiple interpretations of an event, we either force teams to use one number (this rarely works), use a very loose definition of the metric, or recognise that these interpretations represent different metrics — we might call the CRM sales metric ‘Customer Spend’, and the Commercial team may deal with ‘Order Value’.

Starting to define the metrics

A starting point for these definitions might be:

Order Value is the value of the physical products ordered by the customer, excluding tax and shipping, recorded on the date that the order for the physical product is placed

Customer Spend is the amount that the customer pays, recorded on the date that the payment is made. This includes all aspects of the customer spend, including shipping costs and taxes.

Selecting a single VERSION of the truth

At the enterprise level, leaders may recognise that the commercial team’s performance against their targets represents the most business-aligned view of sales, and choose to use ‘Order Value’ as the default sales metric.

Deriving definitions

It’s important to recognise that defining these metrics is only the start of the process. The definitions above can be modified and extended easily, by referring to the base definitions — Order Value Inc Tax, and Order Value before Discount make perfect sense and can be defined against the base definition of Order Value.

Reconciling between different definitions — essential, or a waste of time and effort?

Is it possible to reconcile between Order Value and Customer Spend? Yes, in principle it is, but as the definitions diverge, it becomes time consuming and offers very little business benefit. Better to ask — is it necessary to reconcile them? If reconciliation is required, it is much easier to reconcile both back to the source events than to do cross-reconciliation. Different metric definitions can be viewed as ‘reconcilable, but not reconciled’ until it’s absolutely necessary to reconcile them. If we are confident that A reconciles with B, and A reconciles with C, what extra benefit do we get from reconciling B with C?

Why not just use finance or accounting numbers?

Faced with a challenge to the orthodoxy of the existing metrics, it’s a natural response to turn to the team that is already constantly running the numbers, and they are usually happy to help.

Accounting is undoubtedly an essential business function, uses battle-hardened software and is regularly audited by independent third parties. But accounting numbers rarely give a picture that aligns with the needs of other business groups. They may not track individual sales, the rules around how and when to recognise income and expenditure may be arcane and subject to remarkable flexibility, and aspects of data that are indispensable to non-accounting parts of the business may simply not be present, not being relevant to the core accounting functions.

Commercial finance numbers are hopefully closer to the business needs, but may still have chosen business logic that doesn’t align with needs of other teams, or which makes it easier to reconcile with their accounting colleagues — for example only recognising sales when they are shipped, or processing order data when it is received by the platform rather than when the order occurs. We might choose to define this metric as Revenue — a third view of the same event.

The complications of FX rate conversion

Multiple operating currencies

In these increasingly global times, transacting in more than one currency is increasingly likely. This brings additional complications to arriving at a sales metric, because there will always be a need to aggregate sales in different currencies to one or more standardised currencies. Photobox operates websites in eleven different currencies, budgets in two standardised currencies (depending on the business unit) and summarises to the global level in one group currency.

A metric definition should specify the default approach to handling multiple currencies, and whether other approaches will be provided — e.g. ‘Order value will display figures in the business unit currency by default, but will also be available converted to group currency’.

Multiple FX rate approaches

At the same time, in retail it is very common to use multiple parallel FX rates. Retail plans are usually done using a planning rate, which is fixed for the duration of the plan, but FX rates are changing constantly and many teams will prefer to use a more dynamic rate — a monthly averaged rate, an end-of-month or end-of-day rate, or even intra-day rates.

Choosing an FX rate provider

An additional complication is that unlike stocks and shares, there is no single ‘official’ buy and sell (bid and offer) price. Currency traders are free to arrange deals based on their perception of value, so for general business use a trusted source must be identified. This is usually either a central bank (e.g. Bank of England, European Central Bank, Federal Reserve) or a commercial provider (OANDA, WM/Reuters, fixer.io). Central banks generally freely publish end-of-day rates from and to their corresponding currency, whereas commercial providers normally license a much richer variety of rate APIs, usually for a fee.

Currency Triangulation

If a central bank provider is used, it may be necessary to triangulate via that bank’s currency. The ECB only publishes conversion rates to and from Euros (EUR), so to convert from Swiss Francs (CHF) to British Pounds (GBP), we have to first convert from CHF to EUR, then from EUR to GBP.

Be aware that using a provider that offers currency conversion between multiple pairs of currencies, it is possible that triangulating (e.g. CHF→EUR, EUR→GBP) may produce slightly different results than using a direct conversion (CHF→GBP).

FX Dates

Finally, because there are multiple dates associated with an order across its lifecycle, it is important to record which date is used as the date basis for the FX conversion. It is not always the same as the date on which the sale is recorded.

Conversion approaches

This is a somewhat contrived example based on real requirements I’ve seen — thankfully, in a real situation there is usually more alignment between teams than is shown here!

Roundup — checklist for a sales metric definition

This is by no means an exhaustive list but should be a good starting point for defining your sales metric(s):

Stewardship

• What is the primary purpose of the sales metric?
• Who owns the definition?

Date assignment

• Which date is the sale counted on? (e.g. order created / payment succeeded / order confirmed / order shipped / data received)

FX conversions

• What currency to display the sale in (payment? pricing? warehouse? business unit?)
• Which FX rate is used if currency conversion is required (budgeting? daily? monthly average?)
• Who is the provider of the FX rates?
• Is rate triangulation required?
• Which date to use as the basis for the FX Rate (order created date? payment processed date? order shipped date?)

Sales lifecycle

• Which status of the order counts as a completed sale (e.g. does it count as a sale if the factory cancels the order due to supply problems?)
• Do we account for returns as part of sales metrics? If so — returns need their own set of answers to all of these questions

Valuing the sale

• Whether to include sales of vouchers or by vouchers (pick only one, or risk double-counting)
• Whether to base the value on a regional standardised value or on the transaction amount
• Whether to value before or after discounts (as these may be driven by short-term marketing pushes beyond the control of commercial planning)
• Whether to include sales tax
• Whether to include shipping fees
• Whether to include seller’s fees (if selling via third parties)

Conclusion

Hopefully this shows how behind an apparently simple and fundamental metric there can be a complex and subtle set of business logic, and unless those definitions are available, your business teams may not be working with numbers they need. But, by answering the questions laid out here and providing clear and consistent definitions, you can provide a platform for high quality data so they can make well informed decisions.