How startups investigate and scale their online advertising

Owner of SM Commerce and ZASR Digital, Depesh Mandalia is a UK-based entrepreneur creating global impact through his Facebook ads and growth marketing programs for business owners and students across the UK and Europe.

Mandalia’s background includes playing a key role in the growth stages of Lost My Name’s (now Wonderbly) phenomenal journey, building toucanBox’s offering, and marketing for larger companies like Ann Summers and Tesco collectively generating over £100M in sales since 2005 before launching his Facebook ads agency, SM Commerce.

When I joined Lost My Name, the biggest sales channel was re-sellers like Not On The High Street.
There was some testing with direct sales, but nothing phenomenal. For four weeks, we were spinning multiple marketing plates and one of them stuck; that was Facebook advertising.

Within a few months, we found our traction point and from then it was taken off. We started off spending between £10-50 per day on small tests. We scaled our spend daily to the point where we were spending five figures a month within two months on Facebook. That was a phenomenal experience and it took us all by surprise – we went from selling 30,000 books through resellers in 2013 to 10x the next year.

You don’t have to spend a huge amount when you’re starting out.
If you have half a million invested in your company you can afford to spend thousands a day and scale quickly, but it really doesn’t need to be that much.
Five to ten pounds a day can help you learn a tonne of stuff that you can refine and build on. I’ve built a model that allows us to dip ads into the audience to see if there’s a bite, then quickly scale. The model, called Graduation Testing, is a way of using the auction dynamics to test small and scale fast – built for the startups I worked with initially but applies just as well for larger budget advertisers. 

At the very early stages, don’t expect a return on online advertising.
Having a £10 Cost Per Acquisition (CPA) or a certain target will constrain your testing. Whatever the budget, experiment, and be able to take risks. If that brings you zero return, but a tonne of learnings, is that good enough to develop from?
Testing will be hampered if you’re expecting £500 to turn into £1,500. If the product is new or you’ve not marketed on Facebook before, anything between £500-1000 is required to properly stress test everything across the 4 fundamentals of success on Facebook: your product, audience, offer and funnel – what I call the CORE-4 Principle.

For most online businesses paid search, paid social and affiliates are the top channels I’d develop in the beginning.
Beyond that, it’s taking time to think about where the customer will be. For example, insert swaps have been going on for many years. It increases your reach for very little cost. You want to fail fast and fail often early on, to find those winners early and scale.

A deeper understanding of the customer helps dictate what channel you start experimenting with.
As a founder, you’ll often know better than anyone else who your target market is because you’ve come up with that idea for a reason. At the early stages, that can just be a hypothesis.
For me as a marketeer, it’s about testing that hypothesis. Test, learn, and gain traction. I use my 5W Avatar model focussing on: What, Who, Emotional Why, Functional Why, and Why Not of the product or service. It’s a model I’ve built off the back of Maslow’s hierarchy of needs and the Attack/Defence strategy of David Ogilvy which has served us extremely well on Facebook and other marketing platforms.

I’m a big advocate of personas.
At the early stage, it doesn’t have to be super accurate, just get something down, get an idea of who your target market is. Flesh it out as you start to develop. Once you make a few sales, start to survey them; why did you choose our product, what’s your usual activity around the product? You can gain so much information.

Want to find new opportunities? Survey people after they buy.
We look at Net Promoter Score and split that by relationship. If a group is less satisfied, we can look at ways to fix that. We can make sure we have the right value proposition. Maybe it’s the messaging. Asking customers why they bought can be massively useful – for our own jewelry brand we found men were buying a particular range to celebrate anniversaries, so we used that in our marketing to hone down our message to reach the right people at the right time.

North star metrics differ by business.
For one-off transactions, it’ll be CPA and return on ad spend right off the bat. Then lifetime value and virality further along the line. For a subscription business, it’ll take longer to mature where your ROAS may be negative but you build a model to look at how many subscriptions a cohort takes.

Being a subscription business, your north star metric has to be lifetime value.
Your customer’s first purchase is an acquisition cost. I’d measure lifetime value over a 12-month period where possible, or longer, then break that down by channel and other cohorts like age/gender to look at acquisition costs and time to pay back.
The challenge is optimizing your ad spend on something which could take months to bottom out – at one subscription business we found the 48-hour churn rate to be the most predictive measure of LTV.

A common challenge to subscription businesses is always pushing their ROI.
If you have a payback period of five months any marketing investment you make now, if it’s £10 or £100,000, it’s not going to be positive for five months. Managing cash flow and budgeting around that is really difficult but important to be mindful of.

If there’s one thing I could really press home to founders is, if you have product-market fit and you can validate it, it’s time to start bringing in marketing expertise.
You’ve made those learnings, you’ve done the surveys; the more data and learnings you can give to a marketer the better.

Bring on someone that has deeper experience in the sector, all marketers are different. It’s about finding people that can bring their experience in the sector to help you to scale faster than you can do on your own.
It could be a FT person, a freelancer, or an agency – you get what you pay for but also ensure they have the right experience for the path you need to take. Seek advice if you’re unsure on how to proceed.

A lot of businesses are looking at Silicon Valley and growth hacking, but one of the core misconceptions is that you’ll find a silver bullet.
Actually, it’s about fast iterative testing. For example, if we were testing landing pages for dads, mums, and grandparents, we could develop a page for each, but the fastest test is just to do one for mums and, if that works, do one for dads. It’s thinking about the minimum viable product when it comes to testing as well; the faster you test the faster you’re going to learn.

Michael Jordan talks about how many hoops he missed in his career, but he hit the ones that counted.
And that comes from testing and learning. What I try to bring into a marketing ethos is that you’re going to have lots of misses. But when you get those big wins from those learnings that’s when it really takes off.

Facebook is still a primary route to market.
Since I first made a breakthrough in 2014 Facebook advertising still continues to be the primary channel for many startups. But, the biggest downfall comes from assuming it’s going to be easy – because you can literally turn ads on and start running them today.
I built a system called Brand-driven Performance Marketing (
www.BPMmethod.com) to allow businesses to put the correct structure in place required to make Facebook work consistently and profitably.

Diversity at scale.
The final point on scaling for startups is to not rely on any single channel. Whilst Facebook is a highly scalable platform it’s not without its challenges and so once you find traction, reinvest in new traffic opportunities.

Leave a comment

Join The BPM Method Insider Email Newsletter

Helping you achieve more through better advertising