Tamme is an end to end platform for marketplaces
that are hyperlocal and heavily rely on spending where the marketplace has a deficit in supply or demand.
From measuring where to spend down to a zip code level, programmatically creating the ads, buying / adjusting across multiple channels and adjusting using AI and machine learning to choose sources, imagery etc based off what actually gets outcomes
and revenue. Allowing the marketplace to be greatly more competitive with their unit economics than their competitors’.
Tracking & Analytics
Our analytics are designed specifically for double sided marketplaces (web and mobile apps) It allows us to take into account simple things like knowing that two users are connected through interactions. This allows us to build out very refined individual segments.
It also allows us to gain an inherent understanding of what types of entities are most likely to match so that tamme can create segments on likely matches based on user behavior and user intent.
All of this is then brought down to a suburb / postcode sized area, measuring the supply and demand of the marketplace.
Based on the demand and supply in a region tamme then starts to determine where the advertising budget should be spent and on which side of the marketplace. The first pass of this is fairly naïve as it only takes into account the supply and demand.
tamme then looks at each of the geo-cells and uses the available external demographic data as well as previous performance to determine how much it will cost to gain a user in that location and where the largest impact can be made, as well as what the likely market size is in that geocell, i.e. are there going to be returns from spending money in that location or are there no new users to be gained?
Once the budgets have been set for a location tamme has a look at the channels that are attached to a customer's account (prospecting display banners, Facebook, twitter, LinkedIn, retargeting, call to action messages through email, sms and push) and uses our proprietary machine learning model to determine which channel combinations are going to be most effective for that side of the marketplace, in that region.
tamme then takes into account the minimum effective spends that have been observed in each of these channel/region combinations and sets the channel budgets for each side of the marketplace and each region.
Once tamme is aware of where, and through which channels, the budget is to be spent it tailors the advertising creatives accordingly.
This uses a combination of dynamic creative options (such as; "there are 20 bartenders looking for a job in Potts Point right now") as well as detailed element selection, i.e. which images, taglines, layouts have been most effective for converting new customers or moving customers further up the pipeline.
As with every part of tamme all models are trained off interactions in the app, such as ordering a meal, as opposed to clicks on the creative.
Depending on the size of the budget tamme will also run some tests to be able to better train the model and get a better understanding of what causes users of your marketplace to take actions.
Measurement & Learning
Because tamme integrates all the way from your analytics through to the impressions that users see tamme's able to gain a holistic picture of your marketplace.
All success measures are brought back to both moving users through funnel stages and final/repeat goal conversions, such as ordering a meal, booking a car or hiring a staff member.
Then all expenditure is measured back to simple unit economics:
iv. Churn rate
This allows all marketing activities that tamme undertakes to be tied directly back to the core function of the business instead of being measured by vanity marketing metrics.