Discover more from Startup Pirate by Alex Alexakis
Growth Marketing Fundamentals
Introduction to growth marketing, Moosend's exit, 665 jobs, software companies as media companies, Destroy Uber, NFTs, career growth, events & more
👋 Happy Friday! Welcome to Hunting Greek Unicorns #29. I’m Alex, a product guy turned VC, and every two weeks I send out a newsletter with everything you need to know about the Greek startup industry.
📬 If you find this newsletter interesting, consider sharing with your friends or subscribe, if you haven’t already, and join a community of 2,309 Greek startup enthusiasts.
🎙️Growth marketing fundamentals with Konstantinos Giamalis, Chief Product Officer at efood
This week I’m really excited to host Konstantinos Giamalis, Chief Product Officer at efood, on Hunting Greek Unicorns to discuss all things growth marketing. Konstantinos has been part of efood for about 6 years in different product and marketing roles. He has experienced the tremendous rise of the company since the early days to the acquisition from Delivery Hero and now growing the brand across different countries such as Greece, Croatia, Serbia and others. In this conversation we touched upon many interesting areas such as brand as a growth lever, segmentation, A/B testing, organizational changes to support hyper-growth and many more of his lessons learned growing the largest online food delivery service in Greece with more than 13K restaurants at the moment. Read on!
For many teams growth is a blurry area. Is there a right organizational approach if a team wants to focus on optimizing growth? Where should companies start from?
At efood, we usually create a “growth” strategy and apply it across departments. I’ve seen other companies create dedicated growth teams, but in online marketplaces such as ours, growth requires cross-functional efforts from multiple departments to make it happen. In the beginning, growth was mostly impacted by Sales & Marketing, but currently the industry is changing with Product & Business Development getting more active in growth initiatives. Back in the early days, when we introduced efood in a new city we had to align Sales & Marketing teams, so we could solve the “chicken & egg” problem (supply - demand). By the time we launched a new city, if there weren’t enough restaurants in the platform, there were limited options for users to order, which was a red flag for us as the majority orders from specific restaurants, or explores relevant options according to their past behavior. If users don’t find what they want, they won’t order. When restaurants didn’t get orders, they were disappointed and wouldn’t invest in our platform with additional offers and discounts and could eventually abandon our service entirely. This vicious loop can be endless and lead to very bad and irreversible results.
A good way to create alignment between departments is to create and share business OKRs (Objective & Key Results) and assign growth targets to different departments. Sharing targets might be difficult, but sharing objectives is something that can work really well and establish common ground for the involved teams towards a company goal.
What are the most important lessons learned on growth from your journey so far?
I was lucky enough to experiment with a lot of strategies & techniques, both successful and not. Some of the most important lessons are:
Brand as a growth lever
Although “brand” is quite a vague term, investing in your brand can generate significant growth. Brand initiatives can be a branding commercial on TV, a sponsorship in a niche event, merchandising, etc. The two most challenging aspects of such a campaign is targeting and measuring. For efficient targeting you should deeply understand buyer personas. Google Analytics interests provide rich context that you can validate with brand surveys and user testing. Measuring brand is not easy, but some good indicators are i) performance of direct traffic ii) performance of branded campaigns & channels in Google Analytics iii) exit poll questions to users that perform a specific action (e.g. an order). In our exit poll question “How did you find us?, “A friend of mine introduced efood” is consistently in the top, which is a good indicator of brand health & retention. Last but not least, branded performance can have a massive impact on Google Ads. A user that is looking for “souvlakia delivery efood” has a higher conversion rate and significantly lower cost than a user looking for “souvlakia delivery”. Segment your brand and generic campaigns, and cater them accordingly. There is no right or wrong answer on how aggressive to be on generic campaigns, compared to brand campaigns.
Segment or die
Not all users are the same. An old mantra in Marketing says that if you target everyone, you target no one. You shouldn’t have a single strategy for all users. Develop the right buckets, and apply different tactics and budget. If you don’t know where to start, try to create segments with an RFM analysis. RFM stands for recency (ordered recently), frequency (order frequently) and monetary value (high value order). By the time you have RFM in place, you can start analyzing how different dimensions or actions are affecting RFM of users. Do this with the most important actions to have a clear understanding of how users shape their behavior in your platform.
Example: Do users that order pizza generate better RFM scores? If yes, we should encourage those that haven't ordered pizza before to complete their first order from a pizza restaurant with high quality of service and retention rates. To bind it with the first question regarding silos, metrics & departments, if this was an OKR it could be like:
Objective: Users perform their first pizza order
And it will break down to the different departments in the following manner:
Sales: acquire more pizza shops | Marketing: promote pizza shops to users that haven't ordered pizza yet | Product team: promote pizza shops in shops ranking to users that haven’t ordered pizza yet | BI: identify & tag pizza shops that would be ideal candidates for first pizza order (eg. shops that have low minimum order value & high conversion rate) | Business Development: create a partnership with a big pizza chain and create promo offers for users that want to order pizza for the first time.
Regarding the number of buckets or segmentations, there’s no right or wrong answer. You can try k-means to find the perfect number or define segments according to their density, so you can utilize them in Facebook lookalike segments as well. Besides users, not all marketing channels are the same and their performance varies across industries. For example, Facebook would be great for a clothing brand, while LinkedIn would be better for B2B SaaS. There’s a great article from Avinash Kaushik on a framework that we use for marketing channels called “See - Think - Do - Care”. A few generally good tips are:
Use google display network & social media (push marketing) to reach new users. GDN is cheap and has vast reach, so you can use this to promote informative content to users that don’t know anything about your industry, while social media can be effective as well. The goal is to get them on your website to explore your service and unlock the ability to do remarketing on them later i.e. encourage users to land on your website, but not necessarily order.
Do remarketing on returning visitors with additional informative content. Create sequencing of informative posts and promote them to users that have checked your website before. Your goal here is to encourage users to perform a desired action such as newsletter subscription or contact.
Maintain 100% impression share in branded keywords in Google Ads and be creative with generic ad campaigns. Be present in every search query that includes your brand, as the user already knows you and is looking for your services. The goal is to get as many orders as possible, since branded search queries are usually cheaper & generate better ROI. Generic ad campaigns can provide incremental orders, although they usually are really expensive.
Create dedicated SEO pages for relevant search queries. Create pages optimized for specific SEO queries and relevant to your business. Booking.com has a landing page for every accommodation type, country, island, city and language for a reason.
Not all marketing campaigns have an objective to generate orders. Every campaign fulfils a specific section of the customer journey and the goal is to help users understand why our service is the best option for their needs. That led us to evaluate campaigns in different ways. I would highly recommend to experiment and evaluate campaigns using different attribution models (rather than the default last-click model in Google Analytics), or create an evaluation method that accommodates both click-based and non-click-based marketing campaigns. There is no right or wrong, try a couple of schemas and keep the one that is more relevant with your business!
Focus on your core product
By the time you find product market fit, you have to make sure the core product is fully optimized. There will always be new ideas for features, cool additions and fancy things, but teams should not be distracted. The core product needs to be polished without any blockers towards successful transactions. A common trap here is the “happy path” of the user. For instance in efood, a happy path is when a user logs in, adds address, selects restaurant, adds items and pays for the order successfully. Every step of this path, though, contains a significant amount of sub-flows that might not be optimal. A good example here is how we handle closed restaurants. Many customers land on closed restaurants during the day and we have to encourage them to select another available restaurant and complete a transaction, even though their favorite one is closed.
Some tips on optimizing the core product are:
Establish a robust tracking foundation. We use Google Analytics as the primary source of behavioral analysis, along with other sources of qualitative & quantitative feedback. In the case above, first we need to know how many users land on closed restaurants, then identify where they came from (if there are marketing campaigns that drive people to closed restaurants, we should re-evaluate them) and also track which restaurants are impacted the most (big chain vs. local).
Understand every user's pain-flow. Put yourself in the users’ shoes and try to identify every potential pain-flow, understanding how to solve it. So in our case, what would you like to see instead of a closed restaurant?
Experiment. Try to A/B test the suggestions to make sure the best available solution is implemented. For example, how many restaurants should we display? What kind of restaurants? Which restaurant should be displayed first? etc.
Don’t fall for the 9X effect trap. Companies tend to believe that building new features and embracing innovation will always generate more value for users, while users are reluctant to part with what they already have. This conflict results in a mismatch of nine to one, between what companies believe consumers want and what they truly do.
Where do you start from when deciding to run a growth experiment and what are the best practices you follow?
Our growth framework is inspired by Brian Balfour’s growth engine and testing is one of the five pillars. We did a lot of tests in all marketing channels and our product. There are a couple of A/B test examples that are my favorite (here) but overall here are my recommendations:
Move beyond basics in product A/B testing. Go beyond the basics and experiment more by changing UI elements or your product flow. Simplifying your checkout or account page can have a massive impact on your conversion rates, rather than changing colors of a promo banner.
We are a business, not a laboratory. We have a platform that is 24/7 live, has bugs, multiple deployments every day and not everything is perfect. When defining a test there are two ways to do so: the efficient, where the most important variables are set and the perfect, where all variables are set. We prefer the first option, as in case you identify the most important variables, you won’t be generating random results. Use your time wisely, set the most important variables for the test and let it roll!
Don’t leave evaluation to the computer. Most people that A/B test are particularly interested in generating ideas and prioritizing tests. Although this is not wrong, we need to make sure we understand the evaluation mechanism of the A/B test. Depending on the tool (we use an internal one), sometimes you should implement a different evaluation method for each test according to its severity. For reference, you can check the Bayesian & Frequentist approach.
All tests are valuable. Even a negative test or a test that didn’t provide neither positive nor negative impact are valuable. Sometimes we increase our performance, sometimes we learn. Also, there are times when we implement a test with neutral results, because testing can be used as a sanity check for a new feature, which will provide value in the long term.
95% confidence level is not a holy grail. There is no default confidence level. 95% is an industry best practice. Nevertheless, there are times when even Amazon is evaluating its A/B tests with 70% of confidence level. This is highly correlated with the evaluation mechanism that I described above.
It's ok not to run a lot of A/B tests all the time. There are times that we can’t do A/B tests when we launch new features and follow the path of phased release due to time constraints. As I said before, business is an alive and complex organism and we need to adapt our way of work, in order to be efficient. Just make sure to have the foundation to maintain a healthy cadence of testing.
You don’t have to do A/B tests per se. Testing is broader than A/B testing. At efood we literally tested every app advertising network that operates in Greece, we tested programmatic advertising, we did several voucher campaigns as tests and we spent budget on testing unconventional campaign objectives, targeting and bidding strategies in Facebook and Google Ads. As I said before, either you increase your performance, or you’ll learn.
If you are interested to learn more about our Growth framework, I have a dedicated section in my book “Agile Marketing”.
What is the most important skill for a growth professional?
First of all, every modern growth professional should be competent at growth fundamentals. There’s literally a ton of free resources, strategies & tactics out there. Frameworks, metrics, tools, everything. For me the most important skill to develop is business acumen: understand how the business works, how it generates revenue, what are the opportunities & threats of the industry, what are the core problems it solves and why users are using it, which attributional model is more relevant, identify the potential market fit and use the right growth strategy to unlock and amplify it. Every decision should be based on your business model and industry dynamics. Unless you deeply understand them, there is a high risk you don’t implement the best possible solution for your context and just apply what you already know plus some best practices (gentle expression of copying competitors and industry leaders). This is known as the “law of the instrument” and it won’t help you or the business to move forward. Understanding the business in depth is crucial to prioritize the efforts, budget efficiently and lead the business & growth towards the right direction.
A nice example is QCommerce or quick commerce, where users are able to order their groceries, etc. conveniently & fast. We are transforming our business model to accommodate Qcommerce although it has a relatively smaller share than food orders, and lower margins than a pure food delivery marketplace. Our job is not to deny the initiative, due to lower performance (in the short-term), but to optimize and make it work because it's the future and will help us better establish our market position.
🦄 Startup Jobs
👉 The Greek startup industry is heating up! If you’re still watching from the sidelines, start your job search here. A list with 665 handpicked opportunities waiting for you to apply. Company data is also available.
Moosend, the email marketing & marketing automation platform, was acquired by
Sitecore, a customer experience management company that has raised over $1B from EQT Group.
One of the most interesting personalised healthcare products out there, driven by big data and ML, is ZOE. They just topped up their Series B round with $20M, bringing the total raised to $53M.
iSIZE, a startup that applies deep learning to optimize video streaming & delivery, raised a $6.3M round led by Octopus Ventures.
Emvolon, developing a platform for distributed chemical production from resources that otherwise would be wasted, announced a $1.5M seed investment.
The 3rd cohort of the Visa Innovation Program was announced with 11 fintech startups joining, including: Finloup, Parkadoro, Roadcube, Wealthyhood, Welcome, Woli.
Skroutz is working on a delivery subscription program called Skroutz Plus (free delivery service with annual fee), which will be live in the next few months, according to this article.
Blueground launched Blueground Pass, a subscription program to move across its apartments in 14 cities in a flexible, subscription-like manner.
🤓 Interesting Reads
Every software company is a media company. Build an audience, own distribution and become your own media company. Check out the post I wrote on Twitter!
Eastern & Southern European B2B startups generate ~5x more revenues per $ invested vs. their US counterparts! Meanwhile, the growth rate of funding is much higher across Europe vs. US. The current upside in countries such as Greece, Poland, Romania, etc is massive. More details here.
A post on career growth by Ioannis Papapanagiotou, Senior Engineering Manager at Snowflake. Growth via an increase of influence & impact instead of focus on career ladder, the Engineering/Management transition, setting career goals & more.
Nikos Drandakis, founder & CEO of Sync, reflecting on his Taxibeat days, how the team managed to “Destroy Uber” in Peru after making it their motto and being laser focused on the one goal that matters here.
Tanya Sharma, co-founder & CPO of Wonderpath, on co-founder relations and communication, commitment and complementary skills as important elements here.
Odysseas Lamtzidis, Developer Relations at Netdata, wrote a post on best practices to optimize infrastructure monitoring within DevOps teams.
New project from University of Nicosia researchers with the aim to calculate NFT market caps, called NFT Valuations, was recently launched.
Argyris Kaninis and Marios Stavropoulos, founders of Softomotive, discussing about the path that led to the acquisition from Microsoft, the transition and latest software they have been developing for Microsoft in Greece and more.
John Short, Founder of Compound Growth Marketing, on brands vs. individuals when it comes to creating content and posting on social media here.
A podcast with Steven Galanis, CEO & co-founder of Cameo, on his journey founding Cameo, lessons learned on culture, hiring, and more.
Pavlos Mitsoulis Dobosz, Staff Data Scientist at Expedia, on what MLOps is about and why it's important in order to scale ML at your organization here.
Some interesting events coming up from the Greek tech community:
“Ρόλοι και επίπεδα σε εταιρείες” by Μικρή Κουβέντα on May 14.
“Greece’s diaspora strategy: how Greece can tap the diaspora opportunity” by Delphi Economic Forum on May 14.
“The Greek Cryptocurrency Community meets the Ethereum Foundation” by Greek Cryptocurrency Community on May 15.
“Using Third Party Resources via AWS CDK” by AWS User Group Greece on May 19.
“Applying ML on graph-structured data - an introduction to Graph Neural Networks” by PyData Cyprus on May 20.
“Bridges with other departments: a Transifex Product team story” by Product Community Greece on May 25.
“Is Athens the new Berlin?” by Brighteye Ventures on May 26.
“Blockchain Analytics” by Business Intelligence and Analytics Athens on May 27.
Stay safe and sane,
Greek Startup Pirate 👋