As the quest for better customer dialogue gains pace, user generated content has emerged as a key focus for 2017.  Will it deliver on its promise or is there a very real danger of it muddying an already confused picture?

Brands understand that consumers no longer rely just on the information they generate. Above-the-line advertising is an important tool but it’s a starting point; a stepping stone for today’s more complex interactions.

Consumers want to be part of the conversation; they want to connect with other consumers as well as with brands; and they’ll use and create different platforms to champion, critique and report on products, companies and causes.

A new era for content marketing

In the retail space, online and mobile channels have combined with in-store experiences to create the ultimate connected journey.

Similarly, banks are using branches to greater effect and blending them with digital services to generate new offerings.

>See also: Retail: the next big industry impacted by AI

More touchpoints mean more opportunity for interactions but this omnichannel world is a data-hungry place. Brands must have more content to fuel the growing number of conversations, which demands a lot of resources.

However, it is not only about quantity – they need better quality content, suited to different channels, which consumers will trust and want to engage with.

User generated content is one answer. Although the material comes from the consumers themselves, the best examples involve some degree of brand collaboration.

It is already established as a critical part of the communications mix but, over the next 12 months, it’s set to take on even greater importance. But it’s not without risks.

Unharnessed and mismanaged, it could have a damaging effect on brands as well as the end-users.

Validity, relevance and sentiment

Consumers influence each other so user generated content is valued highly. However, the sheer volume of information means there is a lot of white noise out there.

How can consumers separate the useful from the inane, the trustworthy from the misleading? Authenticity is crucial and brands have an important part to play in helping consumers identify the most valuable sources.

>See also: Why machine learning will impact, but not take, your job

This is particularly the case with customer reviews. Researching a holiday, technology device, financial service or any other product often involves scouring hundreds if not thousands of pieces of feedback on multiple websites, such as TripAdvisor.

Validity is perhaps the first concern. Most reviews on review websites might well be from real customers but how can you be certain?

Last year, Amazon filed a third lawsuit in its ongoing fight against allegedly fake reviewers. It named sights such as PaidBookReviews.org, which offered positive reviews on the e-retailer’s site.

In June 2015, the BBC investigated the global trade in fake reviews and reported that at least 20% of online reviews were bogus. The obvious loss of trust is damaging to brands, which are having to hone ways to help consumers cut through the masses and present an accurate picture.

If they can guarantee reviews are real, both the positive and negative, they’ll earn greater consumer loyalty and respect. ISO standards are making this assurance of accuracy mandatory – but the innovative companies have already got a significant head start.

Relevance is another important challenge. Consumers must browse through tens, maybe hundreds, of reviews before getting to one that truly speaks to them.

If a forty year old man is looking for a holiday villa for him and his family, how helpful are the reviews from teenagers or retired couples? They want advice from other genuine people who are just like them.

The task for brands is to deliver that granularity and relevance to each specific audience, every time.

A third concern is sentiment. Summarising a four-out-of-five-star rating is helpful but has its limitations.

There is a huge opportunity for brands that can deliver an accurate “emotion factor”, delving into thousands of reviews to pull out and aggregate a host of features, sentiments and ratings.

A simple Google search for Bose QC35 wireless headphones reveals 459 reviews on Amazon, 640 on John Lewis, 165 on Currys and 103 on the Bose website – and that’s just the beginning.

The goal has to be to present the good and the bad in a way that removes friction in the research process and makes the purchasing decision easier and quicker.

>See also: Big data predictions: what does 2017 have in store?

Natural language processing (NLP) technologies have made some headway. Organisations are using it to analyse raw text in reviews to categorise the information and derive summaries.

They can deliver a snapshot of a customer’s narrative but often fall short of delivering bespoke tangible insights. The future will inevitably see brands pair these with advanced machine learning techniques to deliver more practical sentiment analyses.

Insight to power innovation

Perhaps the real value in user generated content comes from how brands themselves are harnessing it. It offers a mine of information to inform the business but, with ever-growing volumes, it also carries a huge risk – have brands simply become data rich, insight poor?

If companies can’t grapple with high volumes, they could lose out on important trends or actually arrive at misleading or inaccurate assumptions.

As brands engage across social, mobile, in-store and elsewhere, they are having to connect insights across channels in new and more innovative ways.

Those that can break the customer journey down into granular segments can gain greater intelligence about customer behaviours as well the business.

Campaign’s Revolution Online Retail Report hammers home the reality that consumers want a blend of offline and online touchpoints.

Notably, Christmas 2016 TV ads for traditionally bricks-and-mortar retailers PC World, HMC and Dominos directed consumers to web and mobile services, connecting them across channels.

‘Click and collect’, Ocado’s one-hour delivery service and HMV’s Txt2buy are all signs of a customer-base demanding greater choice now – and brands responding with greater innovation.

These touchpoints join to create a journey map – and customer reviews are important footprints along the way, enabling brands to understand what affected positive or negative sentiments. What initiatives or products led to what behaviours? Which factors had the greatest influence on the likelihood to recommend? Which errors had the biggest impact and what were the best resolutions?

>See also: How companies must adapt to the digital revolution

The well-documented Volkswagen scandal in 2015 is an interesting example of the power of consumer conversations. When it was reported that many vehicles had equipment able to falsely improve the performance of diesel engines on emissions tests, the manufacturer dealt with a huge PR fallout.

The Harvard Business Review’s analysis of Twitter activity in the months that followed showed how sentiment shifted and became more neutral as the manufacturer initiated recovery efforts.

Immediate access to these types of changes in sentiment would mean brands can react quickly, altering their response or opting to stay on course. Advancements in machine learning is making this more achievable.

Amazon introduced machine learning in 2015 to, among others, improve its five-star rating system.

Instead of relying on pure averages, it updated the rating to reflect the present product experience based on recent verified reviews.

This is just one step and offers a glimpse at how much more could be achieved. Ultimately, by knowing which consumers have a propensity to buy and to recommend, over time, brands can market to them in a far more personalised way.

Machine learning will inevitably pave the way for these more bespoke, individual interactions in the future. Work has begun and all eyes are on which brands will develop, adopt and champion the next wave of innovation.

Sourced by Andrew Mabbutt, CEO at Feefo