Key things that I missed out while running influencer campaign for a brand

Unravel Media April 8, 2022

When I joined Unravel Media about 9 months back and started reaching out to potential customers, most of them were experimenting with the idea of influencer marketing. This naturally pushed me to start thinking about key opinion leader (KOL) marketing which was already happening for many years globally and more in the geos where there are adoption problem of some of the conventional media.

I started to wonder how can I also do this for the brands and differentiate our offering and club this with my prior experience of managing performance marketing for some of the large brands that I helped in the past.

I was in a territory which was very unfamiliar and I saw few posts from brands on LinkedIn where they were asking all sorts of questions to media managers such as do you have more micro influencers or mega influencers? What is the collective reach you have on your empaneled influencers? Can you get reach in certain cities & states in India? There were few questions around commercial model and all of them seem interesting.

Having worked in the area of consulting, performance marketing and analytics, I was questioning does this bring performance to brands or incremental tangible benefits that brand can benefit from or is the old wine in new bottle to say that I can get you reach from best of publishing house and it is brand’s responsibility to optimize those. I remember some of the large media houses selling impressions (without targeting) and lot of corporate buying those for a premium price.

What made sense in all this that marketing is shifting to more peer to peer and that was already visible on Tiktok and IG Reels. Second observation was people trust people and speak dialect. No questions on that.

Okay, what I concluded is that if some way I can present the metrics from influencer marketing in same manner as any other performance campaign metric, I can get rid of all those smart questions and I just have to talk about something that I am good at i.e., performance marketing and CAC and repeat and audience persona. On digging deeper, you can see that outcome of influencer marketing can be presented same way as any performance marketing campaign.

Impression (View)Clicks (Clicks)VisitsInstallRegistrationSaleRepeat PurchaseReturn on Ad Sales

All someone needs to do is to have the right tracking mechanism in place to track these events. Which brought me back to the question of our existence, how do we identify influencers who will bring performance? This has to be celebrity otherwise why would the campaign work?

Not exactly, that’s what I figured in next few months.

As an experiment, we got RO from a brand with a focus of getting users from a particular state in India. A small pilot. The outcome was to be measured in terms of direct performance. No bullshit and our existence was in question!!

If I had gone ahead with selection of an influencer who speaks Hindi and English, their users would have come from everywhere across India and few from outside India and while this is a “good reach”, the end objective of trying from that particular State/Geo would have gone for a toss and we could not have used that for validation or performance would have been scattered across his/her audience base.

So we decided to get influencers who speaks vernacular language, we actually “Geofenced” the campaign to a large extent without drawing any radius targeting. We invented “Unravel Ring of fire”.

Well, fair enough. But I missed out on people who speak Hindi/English in that region/state. That’s a small % which is okay when you are doing a pilot for a brand who is looking to measure your performance and your next RO depends on this. We did okay for the month and missed performance benchmark by about 30%. But what was interesting is that we took a big leap and started thinking how can we go at par rather than thinking this will never work. The brand was generous and liked our thoughts as to how we were able to triangulate pin code level performance using some of the SaaS platforms that we deployed.

Thankfully the campaign continued and we became profitable albeit at a small scale. After couple of months, when we started looking at performance which gets tracked on 3rd party SaaS platform that we use for event tracking, I was amazed to see that there were still significant numbers of incremental new sales were coming from few of those campaigns. As usual I led with skepticism thinking some issue with the tool. That wasn’t the case. The performance was real. But why was this happening? Simple reason that video discovery happened at different time for different people and if they liked what they saw, they just clicked on the link in the description box on YouTube and carried on with their journey. This I had not come across on any other paid media platform that I had worked with in the past. Rather, the moment we switched off the marketing, we would see impact within 7 Days. The discovery came very delayed and hence we termed this as “Unravel’s delayed attribution phenomenon”. This was exciting so I went little far. I drew a cohort to see the incremental sales from the historic campaign and the amazing thing to find was that performance was coming from even 3-4-6 months old campaign. I am showing below the illustration just for reference. The actual data had incremental sales coming from even 9-month-old campaigns.

Since I had data now and enough of it, I could clearly draw few parallels with different channels and could see that the end result measured in terms of CAC for the particular brand, was in sync with any other media campaign that we ran at scale.

Week of MonthDec W1Dec W2Dec W3Dec W4Dec W5Jan W1Jan W2Jan W3Jan W4Jan W5Feb W1Feb W2Feb W3Feb W4Mar W1Mar W2Mar W3
Dec W1XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Dec W2XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Dec W3XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Dec W4XXXXXXXXXXXXXXXXXXXXXXXXXXXX
Dec W5XXXXXXXXXXXXXXXXXXXXXXXXXX
Jan W1XXXXXXXXXXXXXXXXXXXXXXXX
Jan W2XXXXXXXXXXXXXXXXXXXXXX
Jan W3XXXXXXXXXXXXXXXXXXXX
Jan W4XXXXXXXXXXXXXXXXXX
Jan W5XXXXXXXXXXXXXXXX
Feb W1XXXXXXXXXXXXXX
Feb W2XXXXXXXXXXXX
Feb W3XXXXXXXXXX
Feb W4XXXXXXXX
Mar W1XXXXXX

Here is what I learned additionally from the exercise:

  1. Get your tracking right. Otherwise, the brand and performance team will always be debating what’s the outcome of the performance.
  2. Regional creators in most cases worked better than Hindi speaking or English-speaking influencers. Of course, that could be function of product too. Mine was low ARPU consumption product so more generic in nature.
  3. There are ways to do geofencing to some extent on influencer campaign and creators physical territory has much higher performance.
  4. One should always look at affinity categories along with focused category. E.g., if you are running a campaign for fintech brand, you can have some creators from comedy or say Vlog categories to experiment, ofcourse, after looking at age and gender split and more demographic data point if you have those, although, those are difficult to find, we have been able to triangulate some of those.
  5. Engagement rate is a good metric to look at but if you are not selecting right influencer, it will not work.
  6. Subscriber count is good to have but unless you know how fast is the channel reaching milestones on subscriber count, you will have difficulty planning follow up campaign.
  7. A follow up campaign does well if you have selected right influencer or capping a frequency to include mix of new viewers and existing viewer.
  8. If you are retargeting to users who dropped off somewhere in the user journey, you should use similar inventories as the creators. E.g., if you have been using Instagram influencers for your campaign and you want to do remarketing, you should try IG reels with similar CTA so that users can have a recall and connect back to the brand.
  9. For any influencer campaign and genres to perform, if you are only looking at Views or sub count, it is unlikely to work in long run. One should always look at CTR and first product interaction metric which could be sign up for example. If your sign up rates are way below compared to other digital media, it is unlikely to perform. Beyond that, product interaction will take care of the rest of the funnel.
  10. For me, 70% of creators yield very little but the rest 30% of them work wonderfully well to compensate for the lost value.
  11. If I am not using non-performance logic on my other media channels, I lose more. It is essentially same audience persona on a different platform and is likely to react to the messages in a similar manner.
  12. How do you price the media is the most important? There are brands who are paying much higher for audience and there are brands who are paying much smarter. One should price it as per their objective.
  13. Chasing influencer is like chasing price on traded stocks. It is better to stay away for volatility.

If you have any opinion or want to discuss this further, you can write to us – Virtualcoffee@unravel.media

Leave a Reply

Your email address will not be published. Required fields are marked *