Dawn’s October 2025 Future of E-commerce study finds a sector still growing, but operating under mounting margin pressure, shifting search dynamics, and unfinished AI foundations.

Growth endures, but the job is getting harder. Most brands reported revenue growth over the last 12 months, yet 51% say the job has become harder and maintaining margins is more difficult - especially at larger revenue bands.

Margins lead the agenda; data & forecasting are next. Leaders’ top priorities are improving margins, better demand forecasting & inventory planning, and preparing for GEO/agentic experiences, with importance placed on customer experience and reducing costs.

AI is everywhere, but not yet solving the hardest problems. 98% use AI in some form, though usage is concentrated in lighter-weight workflows; the toughest pain points (data analysis, forecasting) remain under-served. Accuracy/hallucinations and ERP/OMS integration are the top blockers. Still, over 90% expect AI budgets to increase and over half call AI expansion a top-tier priority.

Agentic Commerce is coming fast - but playbooks are fuzzy. Most brands expect to be ready within 1-4 years, with premium/luxury moving sooner. Yet readiness is uneven: c. 47% are taking steps, but c. 29% want to but don’t know how. Preparation spans data cleanup, ERP modernization, application revamps, workflow updates, org changes, and training.

Search is in flux; GEO readiness is unclear. 61% are concerned about the changing face of search, split between today’s algorithms and tomorrow’s GEO. Only 14% have a clear GEO plan; many expect to rely on agencies (34%) more than in-house tools (14%).

Analytics & forecasting are the biggest upgrade opportunities. Satisfaction with analytics tooling is low, and switch- readiness is high. Buyers want real-time data and proactive, suggested actions. In forecasting, Excel dominates. 82% see value in better systems, but only 27% will pay more.

Cross-border strategies remain resilient. Despite tariff volatility, many plan to increase their investment in cross-border.

Influencers have become a fundamental marketing strategy. 88% use influencers and 86% see them growing; most manage in-house for control, and use agencies for scale.

Returns is still painful. 60% still find returns problematic, with the push for seamless CX driving fraud.

These findings reflect one of the largest transition periods in e-commerce history. Whilst there is uncertainty, there is also excitement for the opportunity that better systems, data, and strategies can bring to solve persistent pain-points that non-AI solutions haven’t yet been able to solve.

We hope you enjoy reading the findings as much as we did.

Our respondents are evenly split across the US and Europe

We only surveyed decision-makers

We focused our survey on both domestic and international mid-market fashion e-commerce brands

E-commerce continues to grow…

…but 50% say the job is getting harder…

…and margins are becoming harder to maintain

Margins dominate priorities, but improving demand forecasting and preparing for GEO come a close second and third

E-commerce leaders, especially in luxury, have their customers front and centre, prioritising customer acquisition over reducing costs

Lack of automation in data-heavy activities is the main source of pain for e-commerce leaders with operations the least painful…

…and, even though revenue growth and customer experience take priority, when buyers think of improving their technology, internal efficiencies get the highest budget allocation

Drivers of investment in new software

$100-$150m turnover business invest the biggest proportion of their revenue on software beating out larger businesses on total software spend

If ever there was a time to sell software it’s now. Across the board budgets have drastically increased compared to last year

98% of respondents are using AI in some form but most of this is in
informal individual use or across a few pilots

AI is being used over a range of use cases, but there is little overlap between the top three pain points and where AI is being deployed, suggesting adoption may be more palpable in lower stakes or less complex workflows

Over half of the respondents think expanding the use of AI is a top priority…

…and over 90% think AI budgets will increase even if it’s just by a little bit

The main challenge in adopting AI appears to stem from a lack of industry-specific solutions that impact how easily it can be implemented and how accurately it can be used

Considering the challenges in aggregate can obscure the most painful challenges when they do emerge, brand tone in particular is considered much more painful when AI doesn’t get it right

There is a discrepancy between the most painful problems of data analysis and demand forecasting & inventory planning and the use of AI to solve them…

It becomes clear that in both cases where accuracy and data integrations are vital, AI just isn’t accurate enough or easy to deploy yet…

No one knows what it means…
There is no universally agreed on definition of agentic commerce, although themes around autonomous decisioning, personalisation and automation emerge.

…but it’s provocative, it gets the people going

Despite being the least mentioned overall priority (p.8), and even though the definition of Agentic Commerce* is still evolving, most brands (especially premium and luxury brands) think they need to be ready within four years.

*Dawn’s definition of Agentic Commerce: A model of online shopping where AI agents autonomously handle the entire purchasing process on behalf of a user, from product discovery and selection to payment and order management (i.e., AI Assistants shopping on behalf of people)

There is a high degree of optimism for how much time it will take to get ready for Agentic Commerce, with only a few e-commerce leaders thinking it will take more time than they’d like

However, almost a third of respondents say that although they’d like to get ready for Agentic Commerce, they don’t know how

Everything, everywhere, all at once

Getting ready for Agentic Commerce is a full business transformation with a strong focus on data preparation, technology upgrades, and stakeholder management

3 in 5 e-commerce leaders are concerned about the changing face of search

Where leaders are concerned, it’s almost a 50/50 split between today’s changing search algorithms and tomorrow’s GEO

Whilst there is optimism and excitement about the opportunity that a changing search environment presents, there are clear concerns around the impact on visibility, ability, and control

Reasons for concern:

Reasons for lack of concern:

Not everyone is prioritising getting ready for GEO but for those that are,
the path is largely unclear

Given the lack of clarity it’s unsurprising that 34% of respondents will be relying on agencies to get ready for GEO and only 14% will be investing in their own tooling to do it themselves

Almost no one is thrilled by their analytics tooling

Even with 6+ scoring, the vast majority of respondents would switch to a better product

Respondents are looking for a range of new capabilities when looking to switch, but the most important are real time data and proactive insights and suggestions

Missing capabilities in analytics tools used by those open to switching

Almost 90% of leaders think they could do demand forecasting and inventory management better, reflecting its position as the second most painful e-commerce workflow

Excel is the most popular tool by far and for some it’s the most sophisticated set up they know of!

82% think they could benefit from a better forecasting system...

...but only 27% are willing to put in the work and the money to upgrade whilst 53% would be interested in an easy and cheap switch

Despite 65% respondents re-evaluating their cross-border strategy post Trump tariffs, cross-border trade continues its inevitable rise

Both Premium and Luxury brands feel similarly

There appears to be only one exception to the rule for brands with revenues between $150-250m, but still 1⁄4 are planning on expanding their cross-border sales

88% of companies use influencers in some capacity with almost half as part of an ongoing strategy

This trend is going nowhere, with 86% of respondents considering them a growing part of their marketing strategy

Most chose to manage their influencers in-house

On balance, companies want to keep managing their influencers in-house.
Where they do want to work with agencies, it’s for the scale

~60% consider returns painful with the rest accepting their fate

Customer experience is the main driver of pain but the cost of both solutions
and the people needed to manage it are a close second and third

Trying to address the main pain of customer experience is the main driver of increasing returns fraud