Deal attribution is a chronic problem in almost all industries. Its direct fallout is the never-ending fights over credit sharing for deals between sales and marketing in B2B industries.
Let me illustrate the attribution problem with the following example:
You read an advertisement for a product in the newspaper at home. You forget about the product after reaching the office. Then you overhear coworkers talking about it. You Google the product, click thru’ the ad, visit the ecommerce company’s website and buy the product.
Do you credit the sale to Print, Word of Mouth, Paid Search or VAR?
As an aside, newspapers are supposed to be dead, so you might be wondering why I chose a print ad in my example. That’s because, contrary to the popular narrative, print is alive and kicking in India. The country is reportedly the only trillion-dollar economy where print is still growing. According to industry insiders, Indian publishers don’t earn even 5% of their print revenue from digital (although I struggle to reconcile that with ad spend figures which put the share of digital ad revenues at 20% of total ad spend in India). But I digress.
Do you know pubs do not earn even 5% of their print revenue from digital ? I wonder which data tell you they survive because of them . But i complete agree that pubs have to blame themselves for the current state of affairs and sooner they correct better it will be for them . https://t.co/1mPNIEJ4Qn
— Rk Agarwal (@rkatweets) January 24, 2021
Analysts have proposed several models to solve the attribution problem.
The one I like the best is CEB / Gartner Attribution Model, which is illustrated in the following exhibit.
First Click Attribution and Last Click Attribution are called single touch attribution (STA) models since they credit 100% of the sale solely to the first or last touchpoint in the purchase journey respectively. The four models in between are called multi touch attribution (MTA) models since they consider all the touchpoints in the purchase journey while attributing credit for sales.
Last Click Attribution is the probably the most popular attribution model.
I can easily picture the following conversation happening a zillion times in sales and marketing departments all over the world.
The customer bought this product on a mobile phone. We will attribute the sale to Mobile. What can go wrong?
Actually, a lot can go wrong by attributing sales to mobile just because it was the last step in the purchase journey. See Teardown Of Myntra’s App Mantra for one the most glaring bungles resulting from the use of Last Click Attribution in recent times.
But Last Click Attribution is easy to implement. It also appeals to marketers who suffer from “calculitis”. With their limited numeracy skills, many of them dismiss detailed attribution analysis as overthinking. Therefore, notwithstanding its pitfalls, I’m guessing that Last Click Attribution will be the most popular attribution model for the forseeable future.
Followed closely by First Click Attribution.
I expect these two attribution models to be well entrenched in B2C businesses where sales cycles are short and the drivers of sales are few.
Now let’s come to the B2B technology business.
Ticket sizes are high. Sales cycles are long. There are many levers of sales. A prospect may have visited the vendor’s tradeshow booth in January, received an email from the vendor in February, downloaded a white paper from its website in March, consulted an analyst about the vendor in April, met with its sales rep in May, spoken to a reference customer in June, deliberated internally in July, and finally placed the order with its Value Added Reseller in August.
First Click Attribution and Last Click Attribution models don’t capture even half of the purchase journey. While the other four Multi Touch Attribution models described above are more comprehensive, they require the vendor to capture all the steps in the sales / purchase journey and process them via Marketing Automation or CRM systems. Very few of them do that, so these MTA models are of academic interest to all but a few vendors.
Which brings us the original problem of deal attribution between sales and marketing in B2B industries.
Rockstar sales reps are self-confident and always acknowledge the contribution of marketing to the deal in areas like events, email marketing, cold call outreach and lead nurturing.
But they tend to be in the minority.
Most sales reps are miserly about sharing credit and and some tend to attribute the deal to claims like “I know this customer from my last job”, thereby cutting out marketing’s contribution to the sale.
This is a major source of friction between sales and marketing in B2B technology companies. As I highlighted at the end of my blost post Driving Greater Sales Marketing Alignment – Sales POV, credit sharing is the elephant in the room of resolving the seemingly-eternal misalignment between the two functions.
The go-to response of marketing is to introduce attribution models.
Wrong answer.
As we’ve seen above, STA models are meaningless and MTA models are impractical in B2B. Therefore, by introducing attribution models, marketing will come across as being pedantic – and thereby reinforce the American Marketing Association finding that overly academic bent of mind is one of the “7 Big Problems in the Marketing Industry”.
A more pragmatic solution is for marketing to come out of its intellectual bubble and start cultivating relationships with sales. This should include but not be limited to taking sales leaders out for a drink from time to time.
Schmoozing might just work where everything else has failed.
That could be the secret of solving the attribution problem in B2B technology.
As an aside, attribution is hard in many other fields apart from B2B technology. Take healthcare for example.
I recently came across a manifestation of the attribution challenge in an Indian Express op-ed about coronavirus data. The authors found it difficult to accept the low death rates reported in India and insinuated that the authorities were fudging data. I countered back, pointing out that the disconnect could stem from how Covid-19 deaths are attributed.
Attribution of Cause of Death (COD) would play a major role in comparing Covid-19 DPM (Death Per Million) rates across different countries. I recently went through a first hand experience, which exposed a big disconnect in this space. At home, the patient had Breathing Trouble (ARD – Acute Respiratory Distress) and very low oxygen saturation levels (40% SpO2), was taken to Emergency & Casualty department of nearby hospital, diagnosed with LRTI (Lower Respiratorial Tract Infection), treated assuming Covid-19, and died before RT-PCR test came back three days later – as negative. Eventually, the correct cause of death was ESRD (End Stage Renal Disease). But First Click Attribution and provisional Last Click Attribution were Covid-19 (whereas the final Last Click Attribution was non-Covid-19). Based on the provisional cause of death assigned immediately after the death, this was counted as a Covid-19 death. But that was wrong based on full facts that emerged later and amounted to an overcounting of Covid-19 deaths. While the op-ed acknowledges the challenge of attributing COD in the case of comorbidities, it does not provide any resolution. Without resolving the challenge, the study is incomplete and is irresponsible for questioning DPM rates and insinuating authorities of fudging data.
With increasing digitalization and reducing computing costs in today’s world, it should be possible to capture and process more comprehensive data related to different steps of various journeys, and accordingly solve the attribution problem in the forseeable future, whether in marketing or healthcare.
However, growing privacy concerns and data protection regulations like GDPR in Europe and CCPA in California tend to dampen my enthusiasm.
Only time will tell how this will play out.
