In retrospect you can see where the rot started: when we stopped writing down what respondents said and started filming and photographing them. It was but a short step to giving them digital audio recorders, cameras and camcorders and asking them to capture their own lives and reflect on what they took.

Then we were really in trouble because they wouldn't shut up. What had begun as a happy few hours of analysing groups turned into a multimedia nightmare with no end to the richness of the data they were willing to provide. Even worse, lots of them enjoyed advertising and watched documentaries about how to build great brands thus making them marketing savvy. Yet still we thought we could get away with it - even when they asked the dreaded question: "Are you asking if I like the advertising or if I think it's any good?"

Of course, the old school researchers never put up with any of this: 'just ask - interpret - and debrief the client'. As long as nobody was required to go back and check with respondents whether we had fairly represented what they thought, we knew we weren't going to get caught. But as soon as those trendy internet start-ups began to build online research panels our troubles really started. Clients then saw no reason why they shouldn't build their own customer panels, capable of interacting with them 24 hours a day. Yet, damn it, that's exactly why you need the objectivity of the researcher as gatekeeper. Have you any idea how contaminated these respondents are? The longer you interview them the less reliable they are. That's why we catch them fresh every time and throw them straight back in if there's a hint that they've been talking to other researchers. Honest guv!

The world has moved on - at least in terms of trying to find new respondents. Customers (I can't call them consumers any more - because there's nothing passive about them) have never been so determined to engage with the products and services they buy and the companies behind them. Most, if not all, of the new communication channels have interactivity built in - and these are buzzing. Bits of existing channels which still buzz, meanwhile, do so because they have found a way to add in interactivity.

This puts pressure on the conventional research model where we create a theatre set apart magically from everyday life. Are the outputs from this that much superior to channel surfing where our customers talk back to us through diaries, blogs and bulletin boards? As if this were not enough, one of the most startling developments in all this interactivity is that we now have even less idea about who's doing the interacting. Not only do people have different telephone numbers and different email addresses, they can also take on different identities.

We may have had problems teasing apart one need state from another when a shopper was browsing the aisle, but at least it was possible to identify a single individual behind them. Now we have no way of telling fraudsters and fabricators from the fragmented selves feeding back relentlessly.

Marketing think tank Nilewide devoted a whole issue of its weekly marketing briefing recently to virtual worlds and role playing - and believes market research to be one of the major beneficiaries. The amount of time and money taken up by these spaces is now larger than most third world economies. These environments aren't just the provinces of social misfits. When you add in the hobbyists, the celebrity hunters and the reality TV show grazers it would appear that many of our customers are spending significant amounts of their time in these places.

Instead of demanding that they identify themselves and tell the truth, why don't we embrace these alternative realities wholeheartedly? Rik Dragonslayer may find it easier to talk about Clearasil than Ricki who has dreadful acne and thinks he'll never get a girlfriend. Margery (33), meanwhile, may be Margery (b.1950) but more comfortable being marketed to as a 33-year-old than someone aged 55.

In this brave new world of interactivity the researchers who survive and flourish will be those who don't fight a pointless rearguard action to separate process from content and projective exercises from reported reality but who deliver research objectives: insight and actionable findings. They will be better party hosts than gatekeepers, adept at managing interactions rather than keeping party crashers out. They won't insist on doing everything themselves. Market research is also set to mutate in ways unimaginable to us now.

It was never about telling the client the truth any more than advertising was about telling consumers about products. Research is taking its place alongside the other communication channels as a way of creating value out of audiences. But it can no longer claim to be separate from them.