Contacts per 100 orders: measure and reduce your shop's contact rate

The customer contact rate in ecommerce answers a question that a raw conversation count hides: do customers need help more often, or are they simply buying more? When a shop sells twice as much, twice as many questions is not a decline. If orders stay flat while questions grow, something in your offer, delivery or communication has stopped being clear.

This guide walks through four steps: a consistent definition, a reason attached to every case, picking the two biggest reasons and fixing them where the question starts. There is no norm here: range, price and delivery method shift the result so much that your own previous period is the only fair benchmark.

How to calculate contacts per 100 orders

The formula is simple: contacts in a period divided by orders in the same period, multiplied by 100. The hard part is the word "contact". One customer may use chat, email and a callback request for the same issue. Count that as three and the rate rises because of scattered channels, not new problems.

Settle the definition once and note it next to the result. Each decision below changes the number, so switching mid-year breaks comparison with earlier weeks.

↔ Scroll the table sideways to see all columns.

Case counts and order counts come from different systems, so pull both for the same date range and time zone. In Meetlead Analytics you can see conversations and tickets for a chosen date range and export data to CSV or XLSX, but you copy the order count from your shop admin.

Tag every contact with one reason

The rate tells you that something changed, not what. For that, every case needs a reason when it is closed. A short list is enough to start: shipping, return or complaint, product after purchase, address change, payment and "other". How to build and maintain a wider set is covered in the guide to tagging conversations and tickets; here only the "why did the customer get in touch" dimension matters.

• One main reason per case, even when the customer asks about two things.
• The reason is chosen at closing, once you know what the issue really was.
• A short description for each tag so two people tag the same case the same way.
• Pre-purchase questions get their own tag and stay out of the count.
• The "other" tag is reviewed weekly, because new problems show up there first.

In Meetlead, tags have a colour, description and usage counter, and you can filter the inbox by several tags at once. A tag added in a conversation is stored on the customer profile, though, so the counter shows how many people carry it now, not how many cases came in this week. For a weekly count, log the reason of each closed case in a sheet, or record the counter on a fixed weekday and treat the difference as an estimate.

Pick two reasons and park the rest

After two or three weeks, convert each reason to a rate per 100 orders. Do not fix everything at once, or you will not know which change worked. Take the two reasons with the highest rate and read a dozen or so conversations for each.

1. Convert the reasons.For each tag, divide the number of cases by the number of orders and multiply by 100. Sort the reasons from highest to lowest.
2. Read a sample.Open a dozen conversations from the top two reasons. Note what customers literally asked, not your summary.
3. Find the place.Work out where the customer was just before asking: on a product page, in the cart, in a post-purchase email or on the returns page.
4. Name the fix.Record one change, an owner and a date so you can compare the weeks before and after.

Here is a fictional example: a bike accessories shop sees that most cases concern shipping and sizing. Shipping conversations start with "when will you dispatch it?" rather than "where is my parcel?", so the order confirmation is missing a dispatch date. Sizing conversations concern one helmet range that has no size chart. If you cannot tell where a question came from, session recordings, when enabled, show the visitor path before the conversation.

Remove the question where it starts

Most fixes land in one of three places. Product questions call for a more complete product page: dimensions, compatibility, material, box contents. Return questions call for a help article linked where the customer decides. Dispatch questions call for a better order confirmation email; later parcel questions are covered in the guide to "where is my order" questions.

Directive 2011/83/EU (Article 8(7)), applicable since 13 June 2014, requires the trader to give the consumer confirmation of a distance contract on a durable medium within a reasonable time and no later than delivery of the goods. EU countries apply it through national law, so check the rules where you sell. It does not say whether that message must include a dispatch date. This text is not legal advice: EUR-Lex, Directive 2011/83/EU.

Shopify Help describes automated order and shipping notifications as templates a merchant can customise, including with order details. It does not prove that a particular wording will reduce questions; your own tags will show that: Shopify, store notifications.

After the fix, keep the tag and its description unchanged. Compare the same reason per 100 orders for two or three weeks before and after. If the rate has not dropped, go back to the conversations: the information may sit somewhere the customer never visits.

Peak weeks in the fourth quarter

November and December change both the number of orders and their mix: more gifts, more first-time customers, tighter deadlines. So compare a peak week mainly with the same week last year, as long as the definition has not changed. Telling fluctuation from real change is covered in the guide to seasonality in traffic and enquiries.

Before the peak, review last year's reasons and fix the two biggest well in advance. During the peak, watch reasons that grow faster than orders, such as payment questions after a gateway outage or delivery questions before Christmas. Team planning for this period is covered in the guide to Black Friday customer support.

Fewer contacts, but not by hiding the button

There is a very bad way to lower the rate: hide the chat, the form and the email address. Customers do not stop having problems; they post reviews, refuse deliveries or simply do not come back. With every fix, check that contact options are as easy to reach as before. Measuring cases resolved without a conversation is covered in the guide to the self-service rate.

In Meetlead, Help Center articles are available in the widget next to the chat, and visitors can vote on whether an article helped. Published articles also feed Meet AI knowledge. In self-service mode customers can choose the assistant; in hybrid mode it is offered when nobody on the team is available. When it lacks information, it offers to pass the question to your team. Meet AI requires an active application and a plan that includes it.

The contacts per 100 orders rate works best as one of a few numbers in a monthly summary. Its structure is covered in the guide to the customer service report. Add two lines to it: the reason you fixed and its rate before and after the change.

Questions about contacts per 100 orders

What counts as a good result? There is no universal norm. A shop selling made-to-order electronics and a coffee subscription shop will have completely different values. Compare yourself with your own weeks under the same definition, and set targets for specific reasons, not for the overall rate.

Should I count questions asked before purchase? Count them separately. A conversation about choosing a product often ends in an order, so adding it to the count would penalise the shop for helping sell. The rate per 100 orders covers cases about orders that have already been placed.