Done well, original research has always been content gold. You only have to look at the number of benchmark reports, industry indexes, and consumer surveys published every year to know there’s some serious brand awareness and authority to be had, as well as a bumper crop of leads and subscribers with the right gating strategy. 

And then there’s the potential earned media value when other companies, influencers, and content creators quote and reference your findings, extending your brand’s authority to new audiences. 

In the age of AI, original research has become an even more valuable, dare I say essential, tactic in the content marketing toolkit. As we’ve explored elsewhere, the marketing funnel is breaking – and visibility increasingly depends on how AI systems surface and reference content. 

The exciting promise of the internet has always been that you can do your own research. Yet, our ability to find the answers we need has grown incrementally harder.  

Over the years, the sheer volume of content exploded; and not all of it was great. Too often, original thinking, detailed knowledge, and reader value took a back seat to chasing keywords. Meanwhile, PPC placements, featured snippets, knowledge panels, shopping links, and a variety of other elements increasingly cluttered the search results, until users didn’t always know which links had earned their rankings through reputation, relevance and authority, and which had simply paid or optimised their way to the top.  

The signal became buried in noise. 

AI search cuts out the noise.  

In AI search, primary sources win 

AI search doesn’t only differ from traditional search on a technical level, evaluating and surfacing information in new ways. It changes the nature of search at a behavioural level. 

People use search engines to look up information. People use AI to jump straight to the answer.  

I no longer need to seek out and plough through lots of reports in the hope one of them might contain the stat or nugget of information I need for my article. With the right prompt, AI just gives me the nugget, complete with sources and links so I can verify the information myself. 

But while AI can curate and summarise any information currently available online (and accessible to LLMs), it cannot generate new knowledge. 

This is why original research represents such a huge opportunity for brands to boost their visibility in ways that not only cut through the noise but also build authority and trust with an audience actively engaging with the topic. 

As long as the research is original – whether that’s new data, fresh insights, or simply findings that don’t already exist online – you’re providing AI with information it can’t find anywhere else. When users ask relevant queries, the LLM has no choice but to reference your content. Your brand stops consuming or recycling the authority of others and becomes a primary source itself. 

That’s true thought leadership. 

The pitfalls of poor research 

Before we explore how to do original research well, it’s worth being honest about how often it’s done badly. Poor research doesn’t just underperform; it actively damages your brand’s credibility. Unfortunately, there’s a lot of it about. 

Recycled research 

If you’ve ever searched for current stats on a topic, you’ve probably come across high-ranking articles with titles like The Social Media Stats All Marketers Need to Know in 2026.  But when you click on the link, the article turns out to be nothing more than a long list of headline statistics and one-line findings drawn from other people’s research, with perhaps a few lines of generic commentary linking them together. This kind of thing has long been a popular way for brands to rank high for relevant topics. The year in the title and the recent publication date, coupled with the sheer amount of (keyword-rich) information, give the content a deceptive veneer of topical freshness and authoritative depth.  

But when you follow the source links (assuming they’re provided), the stats are usually outdated, contradictory, poorly paraphrased or misinterpreted. It’s extremely poor desk research, absent any quality control. 

You can’t listicle or curate your way into AI citations. Before it generates a response to a query, an LLM searches the web for any new information that might add to or supersede what it already has in its training data. Why would an LLM cite your content quoting someone else’s research when it can cite the original source? 

Poor methodology 

Marketers can still come unstuck when attempting their own research, usually because the methodology is fundamentally flawed, such as a ridiculously small sample size. Original research takes more than a SurveyMonkey subscription. You can’t expect a survey that only polls 35 people to somehow be representative of an entire population. 

Another common mistake is, whether by accident or design, phrasing questions or limiting the response options in such a way that they influence or even force certain answers; effectively push polling. What if the option the respondent wants to select isn’t one of the choices on offer? 

And then there’s the issue of what it is you want the research to show. Too many brands invest in research intended to provide answers to questions they want to know, instead of answering what their audience is actively interested in. 

Unsubstantiated findings 

Most marketers – and I include myself in this – aren’t trained data analysts. But we do know how to spin a narrative or sell an angle. 

Under pressure to find something, anything in the results worth writing about, marketers risk highlighting findings that simply aren’t supported by the numbers. Any hint of a pattern, even if it’s statistically insignificant or within the margin of error, becomes a headline trend, while the slightest correlation is championed as clear proof of causation. The research no longer lets the data speak for itself but moulds and cherry-picks the data to fit a preferred narrative. 

Unexpected results 

One risk to original research is you may end up with answers you didn’t expect – or don’t want. There’s no guarantee you’ll be able to commercialise the insights from a research exercise. Fudging numbers or intentionally putting a bias on your findings is bound to backfire and hurt the credibility of your brand. 

CMOs have two choices when the results don’t match up with expectations:  

  • Abandon the project and eat the wasted investment, or  
  • Embrace an opportunity for thought leadership.  

Consider how an unexpected result could help you unlock a new approach to your market or push you to think differently about a topic. If the results were really unexpected or surprising, might they prompt new lines of enquiry to understand why? 

Doing original research right 

If you want your research to stand out and be noticed, it has to offer something new. Your industry or vertical may already be awash with benchmark reports covering all the obvious topics. 

Spend time identifying what your ideal audience would love to know that isn’t currently satisfied by any of the existing research out there. Find that niche within the topic no one else is covering in depth, a niche you can effectively own. 

While a survey might seem like the easiest approach, with the lowest cost of entry, a different format or methodology might help to differentiate your research from the crowd – market research and interviews, perhaps, or drawing on your own anonymised internal data to reveal valuable trends. 

Whatever methodology and format you choose, you’ll need both the analytical capability to interpret the data accurately and the storytelling capability to make the findings compelling and relatable to your audience. While narrative without data doesn’t hold up, data without narrative doesn’t travel. 

This is why I highly recommend finding a research partner. Solid research requires methodological rigour: 

  • How to define and test hypotheses 
  • How to construct survey questions 
  • How to collect and analyse data 
  • How to present your findings  

Working with a research partner also eliminates the risk of bias that inevitably creeps in when a brand conducts and interprets its own research.  

If your budget is a constraint (when is it not?), partnering with the likes of Deloitte, PwC, or McKinsey might be off the table. Instead, seek out a smaller research company or perhaps even a university and see if you can negotiate a co-branding deal where both partners stand to gain. 

Of course, one other challenge is securing the necessary C-suite support, particularly as a research strategy isn’t likely to deliver a quick ROI. You need to champion the many other ways research will generate value for the business. 

  • Increasing AI citations, backlinks and brand visibility 
  • Capturing useful business intelligence 
  • Boosting your subscriber list 
  • Fuelling PR opportunities 
  • Building trust and authority 

Also, don’t be too ambitious from the outset. Highly successful and mature research programs like the Edelman Trust Barometer didn’t start out that way. Maybe publish just a small study at first to stake your claim on the topic and then gradually build it out over time with each new iteration.  

Be the source people trust 

Yes, producing original research is hard, requiring time, rigour, investment and a long-term commitment. That’s precisely why it’s worth doing. If original research was easy there’d be little commercial advantage to be gained from doing it well. 

Start small, find the question your market is asking, and commit to answering it better than everyone else. 

Anyone can publish an opinion. Not everyone can publish a finding backed by solid data-driven research. 

Looking to publish your own original research but not sure where to start? Drop us a line at SALT. Our team of content strategists, writers and researchers can help you develop fresh, insight-filled content to attract positive attention from both people and LLMs.