{{ searchResult.published_at | date:'d MMMM yyyy' }}

Loading ...
Loading ...

Enter a search term such as “mobile analytics” or browse our content using the filters above.

No_results

That’s not only a poor Scrabble score but we also couldn’t find any results matching “”.
Check your spelling or try broadening your search.

Logo_distressed

Sorry about this, there is a problem with our search at the moment.
Please try again later.

Now that TagMan has been tracking all the activity of some very big clients for a substantial period of time, we can provide some pretty definitive answers about how different campaigns appear in, and contribute to, the path to conversion.

From this data, we have proof that natural search and social media channels are vastly undervalued, while the effect of paid search is overstated... 

There were many points to TagMan but one of the big ones was to enable clients to measure,at long last, the real role of individual digital channels in the complete path to conversion.

By housing all the tags from all your channels in one place, clients would be able to see actually where and how different campaigns appeared in and contributed to the path to conversion.

Then, they’d know the answers to some really important, but very difficult, questions. Here’s two of the biggest:

  1. What’s the real value of paid and natural search in delivering sales, particularly on brand terms?
  2. What sales value do social platforms like Twitter and Facebook deliver?

Now that we have been tracking this for a substantial period of time, we can provide some pretty definitive answers.

There’s no prizes for guessing we are discovering that natural search and social media are chronically under-rewarded. Or that paid search (both brand and generic) appear overly so.

But, demanding numbers-based proof of what common sense already tells us is the yang to digital’s ying of real accountability. So, here it is:

The findings below are from a real client. The client cannot, for obvious reasons, be named; the raw data on which this insight is drawn is theirs. We can tell you it is a large retailer.

The real value of natural search

In simplest terms, on this one brand, over a six-week period, natural search listings on non-brand terms delivered about 14 times more value than they are currently credited with.

We can say this because:

We measure the ‘Conversions’ of every channel, how many times any one channel or campaign is the last click. If credit/commission is based on last click then this is the number of sales with which any one channel is credited. In our example, generic SEO was the last click on 1,663 occasions and so credited with 1,663 sales.

We measure the ‘Assisted Conversions’ of every channel. That is the number of sales in which a channel appeared in the path to conversion, at whatever stage. This gives us a gauge of just how often a channel plays a part – any part – in delivering a sale. In our example, generic SEO appeared in the path to conversion of 78,117 sales.

We measure ‘Attributed Conversions’, the amount of sales in which a channel appears in a path to conversion divided by the total number of events in that path. This is the clever one. This is what happens were we to divide credit equally for any sale between all the events that appeared in the user journey, known as a flat attribution model.

While we might want to weight credit to, say, the event that was last or the event that was first, Attributed Conversions gives us a sense of the ‘influence’ of a channel since, if a channel appears a lot in paths where users are exposed to 20 events, you can say it has less value than one that appears lots in paths where there were only two or three.

So Attributed Conversions divides the credit equally between all channels in any one path to conversion to give us a sense of its true influence. In our example, generic SEO is responsible for 23,923 Attributed Conversions. 

So:

  • Natural search listings on generic terms are credited with 1,663 sales, when they should, if credit was to be shared equally between all events, be credited with 23,923.
  • Generic SEO gets credited for 14 times less sales than it deserves on [a flat attribution model].

By the way, brand term natural search, on the same methodology, is under-valued by a factor of nine.

The real value of paid search

The big question surrounding paid search is often whether it’s worth spending on PPC brand terms, given you know many of those clicks would have gone to your natural listings anyway.

So, quickly, we can tell you that branded paid search is 2.5 times less valuable than currently thought.

That doesn’t answer the question about cannibalisation, you’d have to turn brand PPC off and compare your figures to understand that, but it does tell us it gets 2.5 times more credit for sales than it should if you were to reward equally all channels in the conversion path. But, interestingly, product/generic search terms don’t fair much better.

Here’s the rundown:

  • ‘Conversions’ (amount of times generic paid search is last click):16,579.
  • ‘Assisted Conversions’ (amount of sales in which generic paid search appears in the path to conversion): 31,494.
  • Attributed Conversions (amount of sales in which generic PPC features in the path, divided by the total number of events in each path): 10,345.
  • Therefore, generic paid search gets 1.6 times more credit than it should.

Why is this? The simple answer is that, because paid search is so often the last click that it gets attributed with a greater proportion of sales than channels that appear higher upstream but which are still drivers of sales.

The real value of social media

Oh the big one, what value to sales does social media really deliver? The real value of social is of course the ability to talk with your customers.

It still seems strange that we have to justify it much more than that but, to answer questions like ‘how much should I invest in social media’, hard measures, like how they contribute to sales, have to be taken into account. 

In our client example, we track redirect URLs from links in platforms like Twitter and Facebook. Clicks on links in these places are tracked alongside clicks and views from all ‘traditional’ channels and that tells us that social media delivered the following (over the same six-week period):

  • ‘Conversions’ (last clicks):100.
  • ‘Assisted Conversions’ (number of successful paths to conversion in which it appeared): 2,433.
  • ‘Attributed Conversions’ (number of successful paths to conversion in which it appeared, divided by total number of events): 795.3.
  • Social media gets eight times less credit for its direct contribution to sales than it should.

As you can imagine, the teams at the client in question are aiming to do all manner of things with this information. It’s probably not as simple as just splitting marketing spend by channel according to ‘Attributed Conversions’, but it might just be…

Paul Cook

Published 3 March, 2011 by Paul Cook

Paul Cook, the founder of RedEye and TagMan, is a contributor to Econsultancy.  

28 more posts from this author

Comments (50)

Comment
No-profile-pic
Save or Cancel
Avatar-blank-50x50

Aleksej Heinze

Good study, which supports the findings of a number of other ROI focused discussions.

In particular the findings on Social Media ROI are supporting work done by Sage Pay report - UK E-TAILERS STRUGGLE TO CONVERT WEB HITS INTO SALES last year. So we need to argue our case a bit better for Organic Search and Social Media campaigns.

Aleksej

over 5 years ago

Donal Langan

Donal Langan, SEO Director at Easyroommate Ltd

As pointed out, this is well known so its nice to put some numbers behind the assumed knowledge.
However, as we all know, for statistics to have real value they need to be over time and from a large data set. So I look forward to TagMan publishing this sort of information across different sectors and longer timescales; so we can really use it to support SEO and Social Media investment.

over 5 years ago

Avatar-blank-50x50

Kevin Hillstrom

All good!

Here's a challenge for you. I work in the catalog industry. My industry frequently says that "80% of online orders are driven by catalog mailings". In other words, an online order happens, and it is generated by a customer who received a catalog in the past 30 or past 60 days.

So catalogers do the exact same analysis that is listed in this blog post.

Then my clients chose to do something interesting. They chose a group of customers, and chose to not mail a catalog to them for six months.

Guess what?

70% of the orders did not disappear. They still happened. The original analysis said that the orders happened because of catalog mailings (just like the analysis in this post). And yet, when the catalog was taken away, the order still happened.

This means that the catalog truly had no impact on the order for 70% of customers ... the customer was pre-destined to order, it was the analyst who attributed the catalog to the order that made a mistake.

The same issue holds for search, and for social. If you took search and social away, the order may well still happen, in fact, the majority of orders may still happen.

The key is to identify orders that would still happen, pull them out, then conduct the analysis above. You'll still end up with a positive story. But the story won't look as positive as above.

over 5 years ago

Avatar-blank-50x50

Patrick Altoft

Can somebody turn this data into an infographic so I can put it in proposals and presentations?

over 5 years ago

Avatar-blank-50x50

Rosenstand

Thank you for sharing these findings. They only confirm the "gut feeling" that most of us working with SEO has. Always nice to have some real data to refer to.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi all,

And thanks for your comments.

@Donal - absolutely, aggregating our client data to draw cross-client/campaign insights would of course be very powerful but, again, needs client permission first. A large part of our proposition is - and has to be - that the data is the clients', no-one else's. Having said that we're currently working on a project with the IAB for a path-to-conversion study across our retail clients.

@Kevin - great insight. This data does tell us how often different channels appear in the path to conversion and give a great indication of what's working and how. But, to truly know what their impact is, you have to pull your levers.

On the catalogue piece, our client at Boden Oliver Elliott is plugging in precisely that kind of data into the online path-to-conversion data we provide him, with great insights. He presented on that very thing at our first TagMeet last wek - his presentation is online at SlideShare: http://www.slideshare.net/TagMeet/oliver-elliott-boden-using-tagman-path-to-conversion-data-to-drive-customer-acquisition

@patrick - we'd like to see that too :) visualising this information is a real challenge and there's no substitute for pretty graphs! If we had the time, we'd certainly look to do that ourselves.

And to everyone else, thanks for your input. It's most welcome.

over 5 years ago

Avatar-blank-50x50

Amelia

Interesting stuff... It goes with my 'gut' feeling, but still it's good to see numbers beside assumptions. Have to say though, more data across more clients would be even better.

over 5 years ago

Avatar-blank-50x50

Kerry Dye

It is nice to have real data, because although you know that the sales funnel is not as simple as "click and buy" there isn't normally any other way of measuring it.

over 5 years ago

Avatar-blank-50x50

bushy

Not convinced at all. Social media is good certainly but as an overall strategy not so much.

over 5 years ago

Avatar-blank-50x50

Jim Novo

Following on Kevin's comments, just because an event occurred doesn't mean it had any influence, e.g. a display impression below the fold generates a cookie but the visitor never saw the ad.

Additionally, counting events does not take into account the "power" of any particular kind of event to contribute to a conversion.

So, for example: given social interaction is biased towards people who are *already best customers*, would it be surprising that purchasers would have a lot of social clicks in the stream? Does that mean the social clicks *caused* the purchase? Or were they just part of social interaction, and the purchase a discrete event?

This data set is a great start! But you need controlled testing to prove the actual contribution of any type of event. I'm sure we'll see some of that coming in the future.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi Jim,

Absolutely. Numbers are numbers. But they need stories. Sensible insight only comes from combining both. Still, now we make those numbers available - finally - to clients, they're using them just as they should - by applying rigour, common sense and trying things to make sure their stories AND their numbers add up.

over 5 years ago

Avatar-blank-50x50

mike groves

Great post and interesting data although I believe that there is potentially a flaw. You have effectively used a flat attribution model that values all clicks equally. I suspect the picture would be different with a U shaped attribution model that weighted the first and last clicks more heavily. Or different still with a time weighted model that gives clicks in the distant past less credit for achieving conversion, whereas multiple searches in a short space of time are boosted. The challenge we all face is understanding, developing and applying robust attribution models. At last count I believe there were at least 11 different methods out there!

That said, the study is interesting and certainly further fuels the debate - and I for one would be interested in any further findings that Tagman can share.

over 5 years ago

Avatar-blank-50x50

Ozio Media

The study produced some interesting results. The comment made by Donal Langan was of value and a great idea. More information from various sectors including businesses both large in small over a lengthy period of time would produce better and more accurate results. The information in this study can go a long way in convincing wary CEO’s of incorporating social media into their online marketing plans.

over 5 years ago

John Duffy

John Duffy, Marketing Director at Nemisys

Hi Paul, great piece of research, just the sort of thing that makes Econsultancy worthwhile!

I'd like to ask a specific question re paid search. We've found that when a paid advert runs alongside a top 3 natural result, the natural result receives roughly twice as many click throughs compared to when there is no paid search running.

Does/would your model account for that behaviour?

Thanks, John

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi Mike,

Thanks and you're absolutely right, Attributed Conversions in this example represents a flat attribution model and the debate about whether all clicks (and impressions) are equal, as you point out, is critical.

We're finding that advertisers prefer the U-shaped model you mention - where the first and last clicks are upweighted (particularly the last) with the rest of the credit/commission split equally between all other events between them.

This enables clients to give more credit for introducing a customer to the path to conversion and converting them, while also recognising that other channels play a role on the way. Is that the right choice? Well I'm not sure I even know myself yet and certainly every client looks at it differently. But, what's brilliant is to have - and to be in a position to have - that conversation in the first place.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi John,

In a word, no. Again, we're into the realms of lever pulling. Our data is real-time though so you can use it to see exactly how results are changing based on what you're seeing in search results, meaning you can test and refine based on a complete and 'live' view.

But, search agency Greenlight is using our datafeeds, plugged in to its own systems, to understand specifically the role of search (paid and natural) in user journeys. Andrew Freeman presented the findings at TagMeet last week (on SlideShare here http://www.slideshare.net/TagMeet/andrew-freeman-greenlight-using-tagman-path-to-conversion-data-to-understand-the-role-of-search). It's got some useful insights including that the search journey is much longer and more complex than we might expect and that PPC has a role 'upstream' too...

over 5 years ago

Avatar-blank-50x50

Keith Simpson

Paul
As a fellow RedEye alumni can I say that I think you've just made as big a contribution to bringing sanity to digital marketing as the original introduction of tags to de-bunk false web analtics data.

Whenever there's an oportunity I try to evangalise to clients that they are wasting huge amounts of their shareholders' money on ppc and and poor natural SEO. I cringe when big clients talk about £2m google expenditure and shudder when an SME dismisses the £100 he (and millions of others) are spending a month with no return. It doesn't happen in any other marketing discipline.

What you need to do is gross these findings up to give a national figure for the waste or over-emphasis that's occuring and get a good PR to get coverage in the national media - starting with the FT.

Go for it and give lazy digital marketing (and Google) the wake-up call it deserves. Good luck.

over 5 years ago

Avatar-blank-50x50

Tim Leighton-Boyce

Thank you for getting some real numbers out in the open. If only we could have more of this!

This is very valuable information. But is there a risk that this is suggesting an increasing degree of accuracy, when the proliferation of devices used by many people means that the quality of the data is slipping in the other direction?

Is the Tagman system attempting to stitch together sessions in ways which go beyond the traditional cookie-based approach and link different devices to individual users, please?

Either way, I agree that this insight is more than we had before and we should grab it while we can!

over 5 years ago

Robert Duckers

Robert Duckers, Internet Marketing Professional at BlueGreen Internet Marketing

This is very useful, thanks Paul.

over 5 years ago

Avatar-blank-50x50

Ozgur

Great post. We, greenlight, recommend our clients to use tagman to understand the value of each channel. Especially the attribution models are really useful!

over 5 years ago

Andy Headington

Andy Headington, CEO at Adido LimitedSmall Business

A really interesting post Paul. At Adido we've been running a similar tool to Tagman for some time on some fairly decent traffic websites.

The findings we have vary greatly for each client. Some clients have virtually no client journey prior to conversions; they visit and they convert. While others have up to 20 pre conversion visits.

While the numbers and data you have provided are certainly thought provoking, I have to recommend to all of your users/readers of this post that you really need to understand your own clients & their industires to get the most useful data. At the moment we don't believe that there is a 'standard' out there which people can use for conversion/attribution modelling.

over 5 years ago

Avatar-blank-50x50

chris tremling

hi i don't understand how you work out attributed conversions... please could you explain in laymans terms please? in the seo example you have 78,117 assisted conversions, you divide this by the total number of touchpoints in all of those assisted conversions? How do you get 23,923? Are you dividing assisted conversions by average number of touchpoints in all of those sales?

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi all, sorry for taking a while to come back. So:

@Keith - great plan :) Aggregating/comparing paths to conversion where the mix, use and execution of individual channels is different across each client is obviously the major issue here (as well as client permission to use THEIR data). But we're obviously aware of the importance of this data (as are they).

@Tim. Thanks Tim and a crucial point. We assign global user IDs to give clients a single view of their customers but we're obviously all reliant on cookies (and third-party ones at that) to a large extent. I've blogged before for Econ on those risks (http://econsultancy.com/us/blog/5643-cookies-under-fire-as-regulators-move-in and http://econsultancy.com/us/blog/6576-apple-digital-s-greatest-challenge).

@Gary Of course true. But, we can say that across all our clients (50+ enterprise e-commerce businesses in the US and Europe), the picture looks much the same. Also, when it comes to attribution models, most are thinking along the same lines.

@Chris. Yes :) In the example you've used (generic SEO I think) the average number of events in successful paths to conversion where it appeared was 78,117/23,923 = 3.3. What 'attributed conversions' tells us how many sales a channel would be credited with if the client were working to a flat attribution model (where every marketing event that played a role gets an equal share of the credit).

over 5 years ago

Avatar-blank-50x50

Andy Hopkinson, Industrial Placement - Marketing Communications at Mercedes-Benz UK

This is a great article, thanks Paul.

I've always had a great interest in search, so this was a great read.

I've always believed paid search is severely over-rated. A sort of expensive 'quick fix', so it's good to see that finally there is information that this may be the case.

It would be great if there was research into whether people are more likely to click on a sponsored link, or a natural link. I would never click on a sponsored link myself, as in my experience sponsored links tend to have less quality content, which is generally why the link is paid for, rather than being naturally high.

If anyone knows of any research/studies into this it would be great to know about them.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Thanks Andy,

Let's be careful here. This is not to undervalue the greatness of paid search - just to value it (and its role) correctly. Greenlight, for example, found that paid search has a valuable contribution to make further upstream than just last click. Paid search is a brilliant thing - it's just that the inability to track (and therefore credit) anything other than last click can mean it gets disproportionate credit (and therefore spend).

Sponsored listings btw are very popular with searchers, particularly for commercial (rather than informational) searches. There's been plenty of research on that but not sure where to look immediately. Others might have good suggestions or try SEMPO.

over 5 years ago

Avatar-blank-50x50

Chris Jacob

Historically in Direct Response marketing online, the branding / upper funnel channels and media do miss out on credit with reporting and attribution. This was largely due in the past to technical limitations with tracking, etc.
However this analysis is misleading in the sense that if one was to remove SEM from the marketing picture, what would the ROI be?
The answer lies in that all good marketing online (and offline) is playing a part and it is the company and technology that actually can action that across channels online and not merely report on it with some form of attribution to the touch points, that will truly be able to maximize return on ad spend.
Chris (@cjbizdev)

over 5 years ago

Avatar-blank-50x50

George Phillip

Organic+Social media is always a good mix. PPC does tend to get overlooked for some reason. Great discussion none the less, would comment further but my lunch break is all but over.

over 5 years ago

Avatar-blank-50x50

Logan Lee

Do you have a view on the assessment of most profitable combinations of marketing events across channels that when sold together (possibly in-sequence) drove higher ROI?

I've often wondered if that paints a different picture then the value of any single source based on attributed conversions.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi again,

@Chris - not sure I understand your argument here. Could you clarify? In terms of 'merely reporting' on it, you're right, which is why TagMan enables automated awarding of split levels of commission based on set attribution models (it's the same technology that allows perfect de-duplication (from whence our clients get their most instant ROI)). Doesn't get more actionable than that! But I don't want this to be a discussion about TagMan so best we talk offline about that?

@Logan - great question. Oli at Boden's whole point around path to conversion reporting is using it to identify the COMBINATIONS - not the individual channels - that drive sales & ROI. I posted on that here a short while back (http://econsultancy.com/uk/blog/6764-optimise-paths-to-conversion-not-channels) and there's some great slides in Oli's TagMeet presentation (see link in comments above) on how he's doing just that.

over 5 years ago

Avatar-blank-50x50

Dag Holmboe

Hello, good article.

I take issue with the comment that the real value of social media is the ability to talk to customers. If that was true, how come I don't send out my whole company (exec, mrkg, sales, HR, even devs) to talk to customers?

Historically, when people have trying to figure out the value of social media, they have focused in on one thing, namely sales. However, there are more than one value is social media; sales is one, consumer insight is another, savings in in customer support, trade media mentions, wom are others. All contribute to the total value of social media.

The key is to assign a dollar value to each return and you get the "real" value of social media.

Dag Holmboe
CEO, Klurig Analytics.

over 5 years ago

Avatar-blank-50x50

Richard Mills

Very interesting reading, very good to know that what we all think is actually being played out in the real world in numbers. Would love to know the client as that would add even more weight to it.

Richard Mills

Director
Digistraction.net

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

@Dag - hi, you don't get all the people you mention to talk with your customers?! agree the value of social is in all those things and more.

@Richard - many thanks. Would love to be able to tell you the client too :) But advertiser data being in their technology provider's hands (and the parent companies of those providers) is another of the issues that we set out to address. So wouldn't be quite right if we did share would it?

over 5 years ago

Avatar-blank-50x50

Bob Frady

Hey Paul -

Nice stuff!

Any chance you can also compare the value of paid and organic search to email marketing and online ad campaigns? Some sort of "marketing tool prioritization" supported by data would be outstanding to see.

over 5 years ago

Paul Cook

Paul Cook, Director at NCC Web Performance

Hi Bob,

Absolutely. We track both for this client, including retargeting. Oh - and affiliates too. But, thought I'd focus on the big debates first (though value of display/retargeting and email and affiliates is all big too of course!).

But this data doesn't really help anyone other than the client in question prioritize their channels - you have to look at your own data in the context of your own campaigns and objectives to do that.

over 5 years ago

Avatar-blank-50x50

Benedict Smith

Really interesting article, and finally some decent evidence.
Anyone planning on making this into an infographic?

over 5 years ago

Avatar-blank-50x50

Thomas

Very interesting reading, i will download your guidebook about SEM and see where it takes me, i get back to you.

about 5 years ago

Avatar-blank-50x50

Andy Hopkinson, Industrial Placement - Marketing Communications at Mercedes-Benz UK

Hi Paul,

Following on from my initial query, do you know of any research on the relationship between SEO and paid search and whether people are more likely to click on a sponsored link, or a natural link?

Any info you have would be great, thanks.

about 5 years ago

Avatar-blank-50x50

Flyttefirma

Many thanks for the tips!

about 5 years ago

Avatar-blank-50x50

Flytte

Great reading

about 5 years ago

Avatar-blank-50x50

Flyttebil

thhhxxx

about 5 years ago

Avatar-blank-50x50

flytteforretninger

many thanks for the tip!

about 5 years ago

Avatar-blank-50x50

Flytning

many thx for the help!!!

about 5 years ago

Avatar-blank-50x50

Google Business Page

It's quite interesting to see the concrete facts flow through and not see some theoretical presumably-sales-like rhetoric thrown around for paid search (I know this isn't advertorial). Despite this, my faith in natural searches and social media conversions wouldn't founder. This is a great perspective and it's nice to know the flip side of the coin. Great job man!

about 5 years ago

Avatar-blank-50x50

Flyttefirma

Great reading, thank you for sharing

almost 5 years ago

Avatar-blank-50x50

Dustin Hill

Very interesting article. It is great to see some actual data on the return for investing in social media.

almost 5 years ago

Avatar-blank-50x50

Lager

Very interesting article, its great to see the grat effort your put into it!!!

almost 5 years ago

Avatar-blank-50x50

Scor

I've often wondered if that paints a different picture then the value of any single source based on attributed conversions.

over 3 years ago

Avatar-blank-50x50

Ejendomsmægler

Many thanks for the tips!

over 3 years ago

Avatar-blank-50x50

Glen Michaelsen, Marketing at Muuv

Great article - Thanks man!

about 3 years ago

Dating Dk

Dating Dk, seo at dating

many thx for the help!!!

almost 3 years ago

Comment
No-profile-pic
Save or Cancel
Daily_pulse_signup_wide

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Daily Pulse newsletter. Each weekday, you ll receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.