Mohammed Elsaady
Each year, Dutch municipalities and one-stop shops launch waves of letters, door
knocks, and campaigns, urging homeowners to insulate, go solar, or install heat
pumps. Yet, most efforts fall flat. Fewer than one in five interested homeowners
actually buy, and the reasons remain a mystery. When I began this EngD, I found
operators sitting on years of untouched CRM records and excel files, tracking every
household they had ever contacted. They knew who they called, what was requested,
and who bought. But the real question lingered: what truly sets adapters apart from
everyone else?
That question became my mission. Now, the answer is in action, revealing what
truly motivates homeowners to buy and pinpointing where campaigns actually
make a difference.
I gathered records from three one-stop shops, each with its own business model,
and painstakingly reconciled their systems. By linking every household to Micro
CBS data on households and house types, I built a vivid data picture of what tens of
thousands of homeowners actually did after expressing interest, not just what they
claimed in surveys.
Some of what came out overturns current practice. The house matters more than
the household: construction year and dwelling type predict whether someone buys
far better than age, income, or education. People who already have solar panels are
several times more likely to install a heat pump, which makes half of all Dutch
homeowners a target group nobody is working systematically. A customer who
bought once is nearly four times as likely to buy again as a new lead, and most
operators never contact them again. And the channel that works for one product in
one type of house can lower conversion for another, so treating every campaign the
same wastes money and resources.
Findings on their own do not change how a municipality plans its next campaign, so
I built a machine learning tool that does. It is trained on the same conversion
records and predicts, for any neighborhood in the Netherlands, which measures
have real potential there, what drives that, and which specific houses in the
postcode are the best fit. An officer picks an area, downloads the address list, and it
goes straight into a campaign. What used to be decided on experience is now
decided on evidence. A short presentation of the tool is available at
https://canva.link/qltfbefechw8hg2.
Verbouwstromen already uses the recommendations in its advice to one-stop shops
and municipalities, and the three operators together serve the vast majority of Dutch
municipalities.
I arrived at this EngD with a decade of experience in software, marketing, and data
science. For the first time, I needed all three skills at once, and the result is
something energy transition leaders can open in a browser and put to use today.

