Holiday Shopping 2026: The Numbers to Plan Around

Every October the big forecasters publish their holiday numbers, and every October they get reported as one headline: a trillion dollars, a record, a percentage. For a store owner the headline is the least useful part. What matters is underneath it: when shoppers buy, on what device, how much of it happens online, and what they will want to know before and after they pay. Those are the things that decide how busy your chat window gets and when.

This roundup takes the forecasts that are published and checkable as of 20 October 2026, puts them side by side, and translates each one into what it means for your customer service. It ends with a worked example and a week-by-week countdown. If you want the full operational playbook, our Black Friday 2026 customer support plan is the companion piece: this article is the why, that one is the how.

The 2026 holiday forecasts, side by side

Four published forecasts and one consumer survey. Read the “what it measures” column before comparing totals: each one counts a different slice of retail over a different stretch of time.

The sources: NRF, 15 October 2026; Deloitte forecast, 10 September 2026; Deloitte holiday retail survey, 15 October 2026 (4,000 US consumers, surveyed 26 August to 10 September); and Adobe Digital Insights, 28 September 2026, which is based on Adobe Analytics data from US retail sites. For comparison, NRF says holiday sales grew 3.1% to $972 billion in 2025, so its 2026 forecast is a faster season than last year.

What it means for your customer service: every forecast points the same way, a little more spending than last year. NRF also expects retailers to hire 325,000 to 375,000 seasonal workers, against an estimated 356,000 last year, and it describes retailers as cautious about it. Big chains are not staffing up sharply, and a small shop rarely can. The extra conversations will mostly land on the people you already have.

The dates that shape the season

Thanksgiving falls on 26 November, so the whole peak weekend, Cyber Monday included, sits inside November. That leaves 25 days between Cyber Monday and Christmas, and a short run between the sale ending and the last safe posting dates in mid December.

Online is growing faster than the whole

Deloitte expects e-commerce to grow 7.5% to 8.4% this season, against 4.0% to 4.8% for retail overall. Adobe’s online forecast (up 6.7%) also runs ahead of NRF’s total (up 4.2%). And in Deloitte’s survey, shoppers say they plan to spend 55% of their retail holiday budget online.

What it means: a growing share of the season’s questions arrive as typed messages rather than at a till or on the phone. A shopper in a shop asks the person next to them. A shopper on your website asks your chat window, or leaves. If last November’s chat was busy, plan for this November to be busier, even if your sales forecast is flat.

The peak is five days long, and it peaks in the evening

Adobe forecasts $47.5 billion of online spending in “the five-day period between Thanksgiving and Cyber Monday”, 17.3% of the whole two-month season. Cyber Monday is forecast at $15.1 billion (up 6.2%), Black Friday at $12.9 billion (up 9.2%) and Thanksgiving Day at $6.9 billion (up 8.5%).

Last year’s weekend shows the crowd behind those dollars. NRF and Prosper Insights & Analytics counted a record 202.9 million shoppers from Thanksgiving to Cyber Monday 2025: 134.9 million online (up 9%) and 129.5 million in stores. Cyber Monday drew 75.9 million online shoppers, up from 64.4 million (NRF, 2 December 2025). And the busiest part of Cyber Monday was the evening: Adobe found shoppers spent $16 million a minute between 8 pm and 10 pm (Adobe, 2 December 2025).

What it means: the heaviest hours of the year fall when most small teams have gone home. A question about a discount code at 21:30 on Cyber Monday is a sale that is happening now; an answer at 09:00 on Tuesday arrives after the sale has closed. This is the single strongest reason to have answers in your chat that work without a person. Our guide to answering website visitors after hours covers the evening shift in more detail.

Most of it happens on a phone

Adobe expects mobile devices to account for 57.4% of online holiday spending in 2026, up from 56.4%, and a record 63% on Thanksgiving Day. Salesforce, drawing on data from its commerce platforms, reported that phones drove 70% of online orders during Cyber Week 2025 (Salesforce, 5 December 2025).

What it means: your shipping and returns pages are being read on a five-inch screen, often in a hurry. Long policy pages get skimmed, so shoppers ask instead. Test your chat window on a phone: does it cover the “Add to basket” button, are its answers short enough to read without scrolling, do the links it sends open properly? Shoppers who prefer messaging apps will also ask on WhatsApp if you offer it.

Shoppers started early, and they are counting

Deloitte’s survey found that 22% of shoppers plan to start earlier than last year, 1 in 4 already bought holiday items during June promotions, and 28% expect their holiday budget to be spent by the end of October, the highest share in six years. Four in 10 plan to shop October sales events, nearly as many as the 44% who plan to shop Cyber Monday. Planned shopping activity peaks at 77% in late November and 78% in early December. Average planned spend is $1,654 per household (Deloitte, 15 October 2026).

Last year followed the same pattern. NRF found that 84% of consumers had started their holiday shopping by Thanksgiving weekend, yet 53% of their shopping was still to do.

What it means: the season is not one spike. Questions about sale dates, price matching and “will this go on offer?” are arriving now. The late November peak is about codes, stock and delivery times, and early December is about whether things will arrive by Christmas. Budget-minded shoppers read the small print (minimum spend, exclusions, stacking), so write those rules down first.

December is about shipping cut-offs

The USPS recommends sending by Thursday 17 December for USPS Ground Advantage and First-Class Mail, Friday 18 December for Priority Mail and Saturday 19 December for Priority Mail Express, for delivery in the contiguous US before Christmas Day (USPS holiday shipping dates, read 20 October 2026). Private carriers publish their own dates.

What it means: “If I order today, will it arrive by Christmas?” becomes the most repeated question of December, and the right answer changes every few days. Publish your own cut-offs (a day or two before the carrier’s, to allow for packing), in one dated table on your shipping page, and plan the moment you change each answer. Offer the alternative too: e-gift cards, click and collect if you have a shop, or a printable gift note.

The shop and the screen are one journey

Salesforce predicts that 38% of holiday retail dollars this year will be “hybrid sales”, where digital research or AI tools lead to a purchase in a physical store, and it reports that 79% of in-store shoppers are on their phones while they shop (Salesforce, 20 July 2026). Treat a vendor’s prediction as a direction.

What it means: if you have a shop as well as a website, expect “can I order online and collect today?” and “what time do you close on Christmas Eve?” in your chat. Put holiday opening hours and collection rules on your site in plain words.

January is the returns season

Retailers surveyed by NRF and Happy Returns expected 17% of holiday sales to be returned, in line with earlier years, and estimated that 19.3% of online sales come back. Shoppers care how it goes: 82% call free returns a major factor when they buy, 76% are more likely to choose a return option that gives an instant refund or exchange, and 71% are less likely to shop with a retailer again after a poor returns experience (NRF and Happy Returns, 15 October 2025; the 2026 edition was not out when we checked).

What it means: the second peak starts on 26 December and runs well into January, and it decides whether a holiday customer comes back. Write your holiday returns window now (37% of the retailers NRF surveyed planned to extend theirs for the holidays), say plainly how exchanges, refunds and gift receipts work, and have that answer ready before Christmas. Our guide to handling returns and refund questions in chat goes through the wording.

AI is now part of the shopping trip, and of the support queue

Adobe expects AI traffic to US retail sites, meaning shoppers clicking through from AI chat and browser services, to rise 130% this season. Salesforce predicts that 20% of holiday e-commerce traffic will come from AI chat agents and that 1 in 3 online stores will have their own shopping assistant live by Cyber Week. Deloitte found shoppers who use search, social media and generative AI together plan to spend $2,893 on average, against $1,175 for those who use none of the three.

The support side moved too. During Cyber Week 2025, Salesforce counted 55% more AI agent customer service conversations than the week before, and 70% more actions such as updating delivery addresses and starting returns (Salesforce, 5 December 2025).

What it means: more of your visitors arrive having already asked an AI about you, and they expect the same speed when they reach your site. Two things help. Clear, specific pages (prices, delivery times, returns rules) are what AI search tools quote, as our guide to generative engine optimization explains. The same pages are what your own chat should answer from.

Worked example: one store’s holiday question volume

Juniper & Pine is a homeware shop we made up for this example. It sells online, has two people answering customers between packing orders, and gets about 140 chat conversations in a normal October week. Here is how its season might look if it follows the patterns above. Every count below is an assumption for the arithmetic, not a benchmark.

The five days of Cyber Week bring about 84 conversations a day, four times a normal October day. Here is how those 420 conversations might split by topic, and who could answer each.

The first three rows and the returns row, 336 conversations or 80% of the week, are questions whose answer is already written somewhere: the shipping page, the code rules, the product pages, the returns policy. Say an AI agent loaded with those pages resolves three in four of them. That is 252 conversations answered without anyone, many of them in the evening. The team handles the other 168: the 84 handed over and the 84 that needed a person from the start. That is about 34 a day instead of 84, and they are the conversations where a person makes the difference: the lost parcel, the double charge, the gift that has to arrive on the 23rd.

Change the assumptions and the numbers move, but the shape is common: most of the peak is a handful of repeat questions, and the hard cases deserve your team’s full attention.

A six-week countdown, from today to Cyber Monday

From Tuesday 20 October there are six weeks to Cyber Monday. One job a week is enough.

1. 1

   Week 1, 20 to 25 October: find your top 20. Export last November and December’s conversations and emails. List the 20 questions that came up most, and note the hours they arrived.

2. 2

   Week 2, 26 October to 1 November: write the rules down. Sale dates and times with the time zone, each code’s minimum spend and exclusions, whether codes stack, a dated shipping cut-off table, the holiday returns window, gift cards and gift receipts. Publish them on your site.

3. 3

   Week 3, 2 to 8 November: load the answers. Put the shipping, returns and sale pages into your chat, write an exact answer for each code, and check every answer on a phone.

4. 4

   Week 4, 9 to 15 November: test with real questions. Ask your chat the top 20 as a customer would, with typos. Fix what it gets wrong, then walk one hand-over to a person from start to finish.

5. 5

   Week 5, 16 to 22 November: decide who covers what. A rota for 26 to 30 November that includes the evenings, alerts to your phone or Slack for hand-overs, and the December cut-off and January returns answers written in advance.

6. 6

   Week 6, 23 to 30 November: freeze and watch. Re-read your pages after the final prices go in, stop changing rules on 25 November, then check the chat every morning of the peak and keep a person on hand-overs on Cyber Monday evening.

After the peak, change each cut-off answer the day it passes, and switch the emphasis to returns on 26 December. The Black Friday 2026 customer support plan expands every step: the questions to prepare, sample cut-off tables, what to measure each morning.

Where EzyConn fits

EzyConn’s AI agent answers your customers on your website chat, WhatsApp and email from your own content: the pages you choose from your site, the files you upload, and exact answers you write for things like discount codes, which it uses first. If you sell on WooCommerce, it can also read your catalogue, so product, price and stock questions are answered from the live list. A specific customer’s order can be looked up through an AI Action connected to your own system, from the Starter plan.

When a question goes beyond what you gave it, the agent says so and offers a person. The customer sees “Thanks, I’m connecting you to a support specialist now. Please hold for a moment.” and your team gets the whole conversation in the inbox and in Slack, so nobody asks the customer to start again. Reports show how many conversations the AI handled, how many went to a person, and the topics it answered least, which tells you what to write next.

The website chat goes live in about 3 minutes with one script tag, and there is a step-by-step guide for Shopify stores. The Free plan includes the AI agent, needs no card and does not expire, so you can load your shipping and returns pages and ask it your top 20 questions this week. See how EzyConn works for online stores, put a free AI chatbot on your website, or compare the plans on the pricing page. When you are ready, start free.

Holiday shopping 2026: common questions

Keep reading

• Black Friday 2026 customer support plan, week by week
• How online stores use an AI chatbot
• Answering “where is my order?” in chat
• What a good chatbot hand-over to a person looks like
• EzyConn for e-commerce