Multi Touch Attribution (MTA)
Multi-touch attribution & its significance towards product management in general & product marketing in particular…
The importance of marketing as a function within an org. pretty much stands widely understood as of today given how an increasingly large number of start-ups are looking to invest in building a solid marketing team pretty early. Good marketing could just be the shot in the arm a product needs sometimes owing to how the results start to show-up, the numbers begin to trickle in favour, with of course the product standing testimony to those claims as a precursor to it all.
But here are a few questions:
What does great marketing constitute?
How do you tell mediocre / average apart from great marketing before the results are out?
Is there a way to really cut down on the turnaround time to determine great marketing?
Is there a way to predict the likelihood of success of a marketing channel so as to help one put the money where their mouth is?
Multi-touch Attribution could be the answer...
Definition:
Multi-touch attribution enumerates all touchpoints of the user journey, assigning each of them some credit towards establishing the influence of those individual channels have over the end result / the outcome.
The days when marketing would span multiple channels each taking up their share of time (and rather significant one too) with absolutely no insight & no means to measure their effectiveness until such time that it has run its course are almost behind us today.
Well. Talking of start-ups & how one could be operating over a pretty tight budget & schedule with acute pressure towards reaching the outcomes it would make total sense to have a measure of the channels that project a higher probability of yielding than firing off random shots in all directions.
Think about it. If you have invested in multiple marketing channels like billboard ads, banner ads, YouTube ads, PPCs to get the word out, would it really suffice if you pin on one of the channels that led to conversion. Its only but natural for one to decide where they would have to continue their efforts based on the percentages of conversion they track across cohorts.
But, how is that even going to work if you arrive at something like:
The user clicked on an YouTube advert & landed on the website eventually converting into a sale
(or)
The banner ad was the first thing the user clicked on, which led to the introduction / building awareness about the product
There’s certainly an element of incompleteness in both those cases above, don’t you think?
Yes, the first channel that helped one take notice of the product / building awareness is crucial alright & so is also the last channel that led to conversion / the actual sale. But to pin exclusively on either of these & attribute all the success to it could be rather spurious, for you don’t know how many times the user went back & forth over checking your product, checking it out across sites, reading reviews about it, carefully observing photos uploaded by other users before deciding to hit on BUY.
It could be safe to say that one doesn’t click on (/ reach that decision to) BUY today owing to the habit of lurking for tons of information about the product which is also largely covered under the Information Seeking & Searching Behaviour.
There are many types of MTA models that have been employed world over by product & marketing teams. But none of those models could really be termed fool-proof when it comes to zeroing in on the exact advert / line / event that triggered the decision of a “BUY” which is why it was believed to be absolutely crucial to go about tapping, analysing the strength of each of those touchpoints alongside the probability of their conversions.
Now, here are a few of those MTA models:
First touch - the first ever interface the audience has with the product
Oppportunity - the interim phase in the funnel where the audience has understood something about the product & the marketing has been successful in evincing some kind of an initial interest
Lead - a further step in the funnel where the incumbent has shown solid interest & could be a minor step away from converting
Last touch - the final interface the audience has with the product before conversion
Other - any other phases major / minor as relevant to the case & can’t be bucketed into the 4 phases listed above
But if you observe, the problem still persists & the question “how did someone arrive at a decision to buy?” remains unanswered post all of this individual attribution.
So, can that problem be combated?
Of course. As of today we have the technology (+ AI) that’s capable of plugging all the figures over the variables that are important to us & then rely on the algorithm to predict the other unknowns so as to arrive at the output so desired, which in this case is the channel carrying a higher likelihood of hitting success.
And there are many tools in the market that could help with attribution & here are some common ones that are used across Start-ups (no specific order):
“Algorithms are built using complex statistical modelling & are heavily reliant on ML (machine-learning) to be able to accurately predict the channels marketing teams could concentrate on so as to optimize their efforts”
But do all orgs. have the time, money & bandwidth to hire those teams & invest on the infrastructure needed. Turns out, many still don’t & tend to take the backseat when it comes to investing exclusively on one such initiative.
Any alternatives around conventional ways?
Going conventional & building journey maps could help just as much.
Ok, picture this.
The user has just seen an advert, which they may have (& should have, given the best interest of the org.) seen several times in all likelihood & is just about to initiate the sign-up process. Now, with the account creation happening on one leg over the background, popping a short questionnaire with these set of questions in the foreground could help:
1. Which of these channels did you happened to notice our advert on… (Please tick all channels applicable)
list out all the channels in decreasing order of popularity as options
2. Roughly how many times did you see the advert in the past week?
numerical range as correlative to the run time of the advert
3. Where did you first hear about our product / org.?
which channel was it (online / word-of-mouth)
Based on the quality & quantity of the participation, it could get more clearer as to how most people are finding the product & that could be enough to pin on one of them channels, channelizing all your efforts towards it. One could also glean more insight from this result to help test out the positioning of the product, helping weed out problems, if any.