
The first half of 2020 has been an eventful one for ecommerce. Of course technical innovation has progressed: Shopify launched a marketplace and a plethora of other services, Google relaunched organic product listings and Google Shopping went free, Facebook introduced ‘shops’, and thousands of other changes. But much of the first half of 2020 was shaped by COVID-19.
The effects of Coronavirus on ecommerce are tough to overstate. Tens of thousands of businesses world wide sold online for the first time. Millions of transactions that would normally have taken place in stores shifted online. Ecommerce went from “high priority” for most retail businesses to “THE only priority”. And so, eCommerce boomed almost overnight.
- Some categories of retail jumped from being 10-20% online, to 100% online.
- Entire countries’ retail engines were kept ticking over, thanks to the ability to buy online.
- Some areas of ecommerce have struggled – particularly in some areas of fashion, travel, events.
- For many retailers, the ecommerce channel has allowed them to thrive. From DIY stores, to home exercise businesses, right through to companies like Tesco, who were up more than 90% for ecommerce sales in May 2020 vs the prior year.
- Dropshipping entrepreneurs struggled with (very) long shipping times; the opportunities were huge, though, as people scrambled to buy online.
As we move into the second half of the year, it’s likely more of us will be shopping in stores once again; more of us will go back to working at least part time in offices; some of us may be tightening our belts. But it is unlikely that eCommerce will go back to the level it was at prior to March 2020. The stage is set for eCommerce to grow, led by the market leader, Shopify.
Coveo has identified seven factors influencing ecommerce and a consistent theme was one of technology and specifically AI.
In this post we’ll examine several areas where retailers and eCommerce companies can add focus, in order to succeed in the second half of 2020 and beyond.
Factor 1: Data & AI
This is a bit of a cheat really, as (firstly) it’s two elements, and (secondly) both data and AI have been hyped consistently for at least a decade. There are some differences this year.
‘Data’ is something that’s always talked about as having high importance. In general there’s usually a lot more data around ecommerce than traditional retail, and in theory it’s a competitive advantage. The reason it has a little extra importance now is that many regular business patterns have changed:
- Your customers are behaving differently.
- Your suppliers are behaving differently.
- Your competitors are behaving differently.
For example: You may have budgeted a particular PPC spend per month at the start of the year, and had a reasonable assumption of how many new customers & how much revenue that would bring each month – quite likely on the basis of last year’s numbers plus the uplift you’d expect from your level of growth.
But after panic buying, physical retail being closed for months, customers of varying financial stability, and competitors either leaning into extra demand or pushing efficiency where there is a lack of demand, it’s likely your PPC market has changed hugely. And that same factor applies across every marketing channel.
All patterns have shifted
Whereas many retailers guide their actions and plans on the basis of patterns from the past, all of the patterns have shifted. In some cases, change shifts to a degree that it is difficult even to guide on the basis of what happened a few weeks ago. Therefore focusing on data is the best way to succeed with your trading strategy, your marketing strategy, your buying strategy, your customer experience strategy.
More customers than ever are buying online at the moment. If your competitors are using data to quantify their behaviour and react to it, and you are not, that gives them a great opportunity to acquire new customers at this strange moment & service them long into the future.
Amazon in the west, and JD.com, Alibaba, Aliexpress and others in the east, are built on principles of data driven merchandising, product recommendations, matching customers with the products they are most likely to buy. If you can create the ability to use data more effectively in the short term in a similar way, it protects you from others gathering your customers, and can have a material impact on your likelihood to succeed in the second half of the year.
The second part of this is ‘AI’. AI is one of those nebulous terms that was hyped for many years. A little like ‘the year of mobile’, which was hyped for about a decade before it quickly engulfed everything after everyone had stopped talking about it, ‘the year of AI’ is one of those things that’s been talked about as on the cusp for the last few years.
And, just like ‘the year of mobile’, ‘the year of AI’ (or perhaps ‘the year of Machine Learning’) basically came and engulfed us a few years ago without many really realizing:
- Most display ads you see around the web have been put in front of you in part because of machine learned algorithms.
- Google’s ads engine – the fuel behind the company’s revenue – has moved more and more toward AI tools (whether they function as they’re labelled is another matter)
- Twitter, Facebook, and all of the other major social networks use machine learned systems to decide what you – a reader – will and will not see. We question ‘free will’ generally, but there is no free will on social media: what we see is guided by robots.
- Amazon have managed to build up an ads business from very little to $4bn in the last quarter of 2019, basically through the power of machine learned systems.
In most cases, machine learning is used to carry out relatively simple tasks:
- Showing a customer the product they’re most likely to buy, rather than random items from your 10,000+ SKU catalogue.
- Ranking your category pages to maximise your likely gross margin return.
- Categorizing customer support requests to speed up life for your CS team.
- Estimating propensity to buy, to inform your CRM campaigns.
I.e.: The nebulous, and in some cases meaningless, talk of ‘AI’ has been replaced by (at least for now) its outcome: processes & algorithms built into tools, to quietly improve your results, built on top of most of the major platforms, baked into ecommerce platforms, or enabled by third party vendors that can plug in all manner of straightforward customer experience tweaks directly into your systems.
If you are able to jump past the hype, and implement some of these tools in specific areas, we’re now at a stage where you can achieve the work of groups of people using simple ‘AI’ tools.
Factor 2: Digital skills
Digital skills have changed the world over the last 50 years. Four of the world’s current top 10 richest people made all their money through digital technology (Jeff Bezos, Bill Gates, Larry Ellison, Mark Zuckerberg), and the other six all have a huge reliance on it – from Warren Buffett and his investments to Bernard Arnault and his luxury empire.

Over the last few months, this chart spread around the world:CHART HERE
It shows US ecommerce as a percentage of overall retail sales. As you can see, in a single eight-week period up to the end of May, eCommerce had grown by the same amount as it had done over the entirety of the prior 10 years.
Every day there is a fairly major story in the press extending this trend:
- Inditex, Zara’s parent company, announced they’re closing 16% of stores and focusing that effort and cash on ecommerce.
- DPD – the logistics firm – in the UK announced they’re recruiting 6,000 more people to help with ecommerce delivery demand after.
- The UK Government has announced their plan to upgrade the country’s ‘digital strategy’, and it seems that basically means ‘trying to grow people’s digital skills).
Efforts in the past to ‘grow digital skills’ have often been ‘pushed’, trying to shove a boulder up hill; in this case, the boulder is rolling the other way, and we’re rushing to catch up with it. Across many sectors, ‘digital skills’ will be in greater demand than there is supply, but in retail – where demand for ‘digital commerce’ has grown so rapidly – that shortage is already visible, particularly for experienced people who’ve been there and done that, and understand whether spending a couple of million on a major project will genuinely reap rewards, or simply be a waste of time and resource.
The job market is in a state of flux, with some great people, in unlucky segments, looking around for new roles. And remote working has been normalized. Combine the two of those, and it means businesses can find talent to help them grow further, and if they’re willing to continue in partial remote mode, that talent may be 100 or 1,000 or 10,000 miles from their HQ, rather than within an hour’s commute.
Factor 3: Personalisation
There are some businesses where personalization doesn’t really matter: If you’re only selling a narrow number of SKUs, there’s not much you can effectively personalize. If you’re selling largely via Amazon, or Google Shopping ads, then you need not worry so much about personalization, as their algorithms do it for you.
But, if your inventory is a few hundred products or anything above, matching your customers with the products they’re most likely to buy is likely the simplest method of increasing the chance that they’ll buy from you in any given visit to your website. Also, if you’re a dropshipper looking to not only sell products through Facebook Ads but want to build a BRAND, personalisation is something you should really look upon.
If a site sells 100,000 products, putting 10 products at random in front of a customer gives you a very high chance of showing them items they’re not interested in. Narrowing down to the 10 products they are more likely to buy based on their purchase history, what other users have bought in their context, reduces the likelihood of showing them ‘the wrong’ product to avoid annoying them, and provide them with a chance to buy.
And whereas at one stage, decent personalization tools were confined to larger retailers, today they’re abundant, and simple to use, and simple to launch – simple enough that most ecommerce sites could have something better than their current system within a couple of months, perhaps within weeks.
Factor 4: Responding to changing customer behaviour
As I’ve said throughout, one of the larger changes at present is that customer behaviour has changed a lot vs ‘normal’ times. That manifests itself in many ways, but a few of the important ones can be split into ‘short term’ and ‘long term’ changes:
Short-term, there are many customer behaviour changes. You will know those for your customers better than any article, and working to execute in a way that takes them into account today, and tomorrow, and next week can also bring short term results. Just a few of the more universal changes within ecommerce include:
Category shift
There’re regular shifts in which categories are purchased at the moment. From weights and home exercise equipment and fabric masks and plastic shields… to DIY products. Staying on top of this, and forecasting it as best you can into the future, can bring positive results.
In most sectors, some product categories have ‘trended’, and others fallen flat. Ecommerce has a huge advantage here vs stores, and this is really one of those areas that can have a huge impact: From saving cash spent on product categories that won’t sell in 2020, through to leaning into trends that suddenly spike beyond usual reality, there are constant opportunities here.
Openness to new brands
Where customers are buying products they don’t habitually buy each week or month, and some are switching to online vs offline, there is an opportunity to open them up to your brand – something they’re likely much more receptive to than at a time when they could simply drop into B&Q to pick up one of your competitors’ products.
Returns sensitivity
Many customers don’t want to queue for the post office. On larger white goods, it’s impossible to return some (suppliers simply not accepting). As a result, many customers are much more sensitive to returns policies and service at the moment
Long-term, there are also some big changes. I’d expect they’ll include:
Ecommerce by default
As already mentioned, whereas before someone may have chosen to drive 20 miles to go into John Lewis, for the foreseeable future many will opt for JohnLewis.com instead. I.e., there’s much greater demand for ecommerce services, and that expansion is to an extent among less savvy buyers.
Price sensitivity
There is likely much greater economic uncertainty on the horizon. Many brands are flooded with stock (particularly within fashion), and suppliers themselves are backed up with even more stock. In theory that leads to a ‘perfect storm’ of discounting. For example: Boohoo are currently up to 70% off + an extra 20%; John Lewis are on up to 70% clearance.
These are predictable: As soon as stores closed, you could predict with 10 minutes thought that this would happen. Either retailers would lock up stock till 2021 (expensive from a warehouse point of view), or negotiate with suppliers to take back stock (hard to handle logistically) or would try to spur sales along by discounting at some point.
These are simple examples, but hopefully illustrate that there are predictable changes among the unpredictable at the moment, and there will continue to be for the remainder of 2020. Putting in the time to predict them, and plan around them, puts you in much better stead than drifting with this year’s tide.
Factor 5: Being prepared for recession
There is plenty of talk of ‘choppy waters’ on the horizon, and many businesses have also sadly bumped into rocks already. Ecommerce businesses are not fully protected from this, but there are some elements that provide a little greater resilience naturally:
- Ecommerce businesses tend to focus more toward customer loyalty than wholly offline retailers – building relationships with customers, rather than simply expecting those customers to arrive in store. The ability to forecast purchase, and react when reality differs from your forecast, mean you can steer with much greater certainty.
- Having relationships with customers also allows you to react quicker: If suddenly all of your VIPs are no longer buying, you can make an assumption about what that means and adjust; or also literally phone each one of them up.
- Ecommerce can also run to much shorter timescales. Whereas measuring the effects of a TV campaign on a group of regional stores may take weeks/months to get any meaningful data and book the next round of ad spots, measuring a PPC account’s impact on an ecommerce site is basically instant. This is often taken for granted, but at a time where price sensitivity is a big issue, being on top of all your marketing channels’ spend is essential, and monitoring for changes in ‘new customer conversion rate’, ‘cost per newly acquired customer’, etc to help steer.
- On the supply side, ecommerce businesses can also act with a bit more agility. Stock is much easier to manage in one central warehouse than a group of stores, and replenishment much quicker. This simplifies leaning into particularly high selling categories.
Some of the activity you may look at to ‘protect from recession’ may be obvious – for example buying in lower priced items to account for increased price sensitivity, or working on your fixed cost structure to protect cash in case of unpredicted need. Some of it may be less obvious: testing different models, for example, extending off your primary business.
As one obvious illustration of that – H&M Group largely sell their own brands in-store. A while ago they launched a ‘test’ brand outlet, basically an online TK Maxx competitor. Initially it was only in Sweden, but over the Covid19 lockdown period they’ve also launched it in Germany, Austria, and the Netherlands. You can take a look at Afound.com.
This may be coincidence, or it may be H&M Group realizing that the channel will be flooded with stock that’s been locked up in stores, and a simple way to create value from that is to create a brand outlet, through which they can use many of their pre-existing processes and skills to react to customer demand, and also perhaps clear some of their own stock.
There is still great challenge for most businesses in ecommerce, and some of the bigger ecommerce businesses will benefit the most, but thinking through where you’re at in the context of your goals, and the opportunities available to you among the above five factors, can hopefully bring a slightly less jarring, more predictable, and more positive second half of the year.